argo11 commited on
Commit
782c1bd
·
verified ·
1 Parent(s): c92cc40

0399 upload checkpoint-3000

Browse files
checkpoint-3000/chat_template.jinja ADDED
@@ -0,0 +1,337 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {#-
2
+ In addition to the normal inputs of `messages` and `tools`, this template also accepts the
3
+ following kwargs:
4
+ - "builtin_tools": A list, can contain "browser" and/or "python".
5
+ - "model_identity": A string that optionally describes the model identity.
6
+ - "reasoning_effort": A string that describes the reasoning effort, defaults to "medium".
7
+ #}
8
+
9
+ {#- Tool Definition Rendering ============================================== #}
10
+ {%- macro render_typescript_type(param_spec, required_params, is_nullable=false) -%}
11
+ {%- if param_spec.type == "array" -%}
12
+ {%- if param_spec['items'] -%}
13
+ {%- if param_spec['items']['type'] == "string" -%}
14
+ {{- "string[]" }}
15
+ {%- elif param_spec['items']['type'] == "number" -%}
16
+ {{- "number[]" }}
17
+ {%- elif param_spec['items']['type'] == "integer" -%}
18
+ {{- "number[]" }}
19
+ {%- elif param_spec['items']['type'] == "boolean" -%}
20
+ {{- "boolean[]" }}
21
+ {%- else -%}
22
+ {%- set inner_type = render_typescript_type(param_spec['items'], required_params) -%}
23
+ {%- if inner_type == "object | object" or inner_type|length > 50 -%}
24
+ {{- "any[]" }}
25
+ {%- else -%}
26
+ {{- inner_type + "[]" }}
27
+ {%- endif -%}
28
+ {%- endif -%}
29
+ {%- if param_spec.nullable -%}
30
+ {{- " | null" }}
31
+ {%- endif -%}
32
+ {%- else -%}
33
+ {{- "any[]" }}
34
+ {%- if param_spec.nullable -%}
35
+ {{- " | null" }}
36
+ {%- endif -%}
37
+ {%- endif -%}
38
+ {%- elif param_spec.type is defined and param_spec.type is iterable and param_spec.type is not string and param_spec.type is not mapping and param_spec.type[0] is defined -%}
39
+ {#- Handle array of types like ["object", "object"] from Union[dict, list] #}
40
+ {%- if param_spec.type | length > 1 -%}
41
+ {{- param_spec.type | join(" | ") }}
42
+ {%- else -%}
43
+ {{- param_spec.type[0] }}
44
+ {%- endif -%}
45
+ {%- elif param_spec.oneOf -%}
46
+ {#- Handle oneOf schemas - check for complex unions and fallback to any #}
47
+ {%- set has_object_variants = false -%}
48
+ {%- for variant in param_spec.oneOf -%}
49
+ {%- if variant.type == "object" -%}
50
+ {%- set has_object_variants = true -%}
51
+ {%- endif -%}
52
+ {%- endfor -%}
53
+ {%- if has_object_variants and param_spec.oneOf|length > 1 -%}
54
+ {{- "any" }}
55
+ {%- else -%}
56
+ {%- for variant in param_spec.oneOf -%}
57
+ {{- render_typescript_type(variant, required_params) -}}
58
+ {%- if variant.description %}
59
+ {{- "// " + variant.description }}
60
+ {%- endif -%}
61
+ {%- if variant.default is defined %}
62
+ {{ "// default: " + variant.default|tojson }}
63
+ {%- endif -%}
64
+ {%- if not loop.last %}
65
+ {{- " | " }}
66
+ {% endif -%}
67
+ {%- endfor -%}
68
+ {%- endif -%}
69
+ {%- elif param_spec.type == "string" -%}
70
+ {%- if param_spec.enum -%}
71
+ {{- '"' + param_spec.enum|join('" | "') + '"' -}}
72
+ {%- else -%}
73
+ {{- "string" }}
74
+ {%- if param_spec.nullable %}
75
+ {{- " | null" }}
76
+ {%- endif -%}
77
+ {%- endif -%}
78
+ {%- elif param_spec.type == "number" -%}
79
+ {{- "number" }}
80
+ {%- elif param_spec.type == "integer" -%}
81
+ {{- "number" }}
82
+ {%- elif param_spec.type == "boolean" -%}
83
+ {{- "boolean" }}
84
+
85
+ {%- elif param_spec.type == "object" -%}
86
+ {%- if param_spec.properties -%}
87
+ {{- "{\n" }}
88
+ {%- for prop_name, prop_spec in param_spec.properties.items() -%}
89
+ {{- prop_name -}}
90
+ {%- if prop_name not in (param_spec.required or []) -%}
91
+ {{- "?" }}
92
+ {%- endif -%}
93
+ {{- ": " }}
94
+ {{ render_typescript_type(prop_spec, param_spec.required or []) }}
95
+ {%- if not loop.last -%}
96
+ {{-", " }}
97
+ {%- endif -%}
98
+ {%- endfor -%}
99
+ {{- "}" }}
100
+ {%- else -%}
101
+ {{- "object" }}
102
+ {%- endif -%}
103
+ {%- else -%}
104
+ {{- "any" }}
105
+ {%- endif -%}
106
+ {%- endmacro -%}
107
+
108
+ {%- macro render_tool_namespace(namespace_name, tools) -%}
109
+ {{- "## " + namespace_name + "\n\n" }}
110
+ {{- "namespace " + namespace_name + " {\n\n" }}
111
+ {%- for tool in tools %}
112
+ {%- set tool = tool.function %}
113
+ {{- "// " + tool.description + "\n" }}
114
+ {{- "type "+ tool.name + " = " }}
115
+ {%- if tool.parameters and tool.parameters.properties %}
116
+ {{- "(_: {\n" }}
117
+ {%- for param_name, param_spec in tool.parameters.properties.items() %}
118
+ {%- if param_spec.description %}
119
+ {{- "// " + param_spec.description + "\n" }}
120
+ {%- endif %}
121
+ {{- param_name }}
122
+ {%- if param_name not in (tool.parameters.required or []) -%}
123
+ {{- "?" }}
124
+ {%- endif -%}
125
+ {{- ": " }}
126
+ {{- render_typescript_type(param_spec, tool.parameters.required or []) }}
127
+ {%- if param_spec.default is defined -%}
128
+ {%- if param_spec.enum %}
129
+ {{- ", // default: " + param_spec.default }}
130
+ {%- elif param_spec.oneOf %}
131
+ {{- "// default: " + param_spec.default }}
132
+ {%- else %}
133
+ {{- ", // default: " + param_spec.default|tojson }}
134
+ {%- endif -%}
135
+ {%- endif -%}
136
+ {%- if not loop.last %}
137
+ {{- ",\n" }}
138
+ {%- else %}
139
+ {{- ",\n" }}
140
+ {%- endif -%}
141
+ {%- endfor %}
142
+ {{- "}) => any;\n\n" }}
143
+ {%- else -%}
144
+ {{- "() => any;\n\n" }}
145
+ {%- endif -%}
146
+ {%- endfor %}
147
+ {{- "} // namespace " + namespace_name }}
148
+ {%- endmacro -%}
149
+
150
+ {%- macro render_builtin_tools(browser_tool, python_tool) -%}
151
+ {%- if browser_tool %}
152
+ {{- "## browser\n\n" }}
153
+ {{- "// Tool for browsing.\n" }}
154
+ {{- "// The `cursor` appears in brackets before each browsing display: `[{cursor}]`.\n" }}
155
+ {{- "// Cite information from the tool using the following format:\n" }}
156
+ {{- "// `【{cursor}†L{line_start}(-L{line_end})?】`, for example: `【6†L9-L11】` or `【8†L3】`.\n" }}
157
+ {{- "// Do not quote more than 10 words directly from the tool output.\n" }}
158
+ {{- "// sources=web (default: web)\n" }}
159
+ {{- "namespace browser {\n\n" }}
160
+ {{- "// Searches for information related to `query` and displays `topn` results.\n" }}
161
+ {{- "type search = (_: {\n" }}
162
+ {{- "query: string,\n" }}
163
+ {{- "topn?: number, // default: 10\n" }}
164
+ {{- "source?: string,\n" }}
165
+ {{- "}) => any;\n\n" }}
166
+ {{- "// Opens the link `id` from the page indicated by `cursor` starting at line number `loc`, showing `num_lines` lines.\n" }}
167
+ {{- "// Valid link ids are displayed with the formatting: `【{id}†.*】`.\n" }}
168
+ {{- "// If `cursor` is not provided, the most recent page is implied.\n" }}
169
+ {{- "// If `id` is a string, it is treated as a fully qualified URL associated with `source`.\n" }}
170
+ {{- "// If `loc` is not provided, the viewport will be positioned at the beginning of the document or centered on the most relevant passage, if available.\n" }}
171
+ {{- "// Use this function without `id` to scroll to a new location of an opened page.\n" }}
172
+ {{- "type open = (_: {\n" }}
173
+ {{- "id?: number | string, // default: -1\n" }}
174
+ {{- "cursor?: number, // default: -1\n" }}
175
+ {{- "loc?: number, // default: -1\n" }}
176
+ {{- "num_lines?: number, // default: -1\n" }}
177
+ {{- "view_source?: boolean, // default: false\n" }}
178
+ {{- "source?: string,\n" }}
179
+ {{- "}) => any;\n\n" }}
180
+ {{- "// Finds exact matches of `pattern` in the current page, or the page given by `cursor`.\n" }}
181
+ {{- "type find = (_: {\n" }}
182
+ {{- "pattern: string,\n" }}
183
+ {{- "cursor?: number, // default: -1\n" }}
184
+ {{- "}) => any;\n\n" }}
185
+ {{- "} // namespace browser\n\n" }}
186
+ {%- endif -%}
187
+
188
+ {%- if python_tool %}
189
+ {{- "## python\n\n" }}
190
+ {{- "Use this tool to execute Python code in your chain of thought. The code will not be shown to the user. This tool should be used for internal reasoning, but not for code that is intended to be visible to the user (e.g. when creating plots, tables, or files).\n\n" }}
191
+ {{- "When you send a message containing Python code to python, it will be executed in a stateful Jupyter notebook environment. python will respond with the output of the execution or time out after 120.0 seconds. The drive at '/mnt/data' can be used to save and persist user files. Internet access for this session is UNKNOWN. Depends on the cluster.\n\n" }}
192
+ {%- endif -%}
193
+ {%- endmacro -%}
194
+
195
+ {#- System Message Construction ============================================ #}
196
+ {%- macro build_system_message() -%}
197
+ {%- if model_identity is not defined %}
198
+ {%- set model_identity = "You are LLM-jp-4, a large language model trained by LLM-jp." %}
199
+ {%- endif %}
200
+ {{- model_identity + "\n" -}}
201
+ {% if knowledge_cutoff is not defined %}
202
+ {%- set knowledge_cutoff = "2025-12" %}
203
+ {%- endif %}
204
+ {{- "Knowledge cutoff: " + knowledge_cutoff + "\n" -}}
205
+ {% if conversation_start_date is not defined %}
206
+ {%- set conversation_start_date = strftime_now("%Y-%m-%d") %}
207
+ {%- endif %}
208
+ {{- "Current date: " + conversation_start_date + "\n\n" }}
209
+ {%- if reasoning_effort is not defined %}
210
+ {%- set reasoning_effort = "medium" %}
211
+ {%- endif %}
212
+ {{- "Reasoning: " + reasoning_effort + "\n\n" }}
213
+ {%- if builtin_tools %}
214
+ {{- "# Tools\n\n" }}
215
+ {%- set available_builtin_tools = namespace(browser=false, python=false) %}
216
+ {%- for tool in builtin_tools %}
217
+ {%- if tool == "browser" %}
218
+ {%- set available_builtin_tools.browser = true %}
219
+ {%- elif tool == "python" %}
220
+ {%- set available_builtin_tools.python = true %}
221
+ {%- endif %}
222
+ {%- endfor %}
223
+ {{- render_builtin_tools(available_builtin_tools.browser, available_builtin_tools.python) }}
224
+ {%- endif -%}
225
+ {{- "# Valid channels: analysis, commentary, final. Channel must be included for every message." }}
226
+ {%- if tools -%}
227
+ {{- "\nCalls to these tools must go to the commentary channel: 'functions'." }}
228
+ {%- endif -%}
229
+ {%- endmacro -%}
230
+
231
+ {#- Main Template Logic ================================================= #}
232
+ {#- Set defaults #}
233
+
234
+ {#- Render system message #}
235
+ {{- "<|start|>system<|message|>" }}
236
+ {{- build_system_message() }}
237
+ {{- "<|end|>" }}
238
+
239
+ {#- Extract developer message #}
240
+ {%- if messages[0].role == "developer" or messages[0].role == "system" %}
241
+ {%- set developer_message = messages[0].content %}
242
+ {%- set loop_messages = messages[1:] %}
243
+ {%- else %}
244
+ {%- set developer_message = "" %}
245
+ {%- set loop_messages = messages %}
246
+ {%- endif %}
247
+
248
+ {#- Render developer message #}
249
+ {%- if developer_message or tools %}
250
+ {{- "<|start|>developer<|message|>" }}
251
+ {%- if developer_message %}
252
+ {{- "# Instructions\n\n" }}
253
+ {{- developer_message }}
254
+ {{- "\n\n" }}
255
+ {%- endif %}
256
+ {%- if tools -%}
257
+ {{- "# Tools\n\n" }}
258
+ {{- render_tool_namespace("functions", tools) }}
259
+ {%- endif -%}
260
+ {{- "<|end|>" }}
261
+ {%- endif %}
262
+
263
+ {#- Render messages #}
264
+ {%- set last_tool_call = namespace(name=none) %}
265
+ {%- for message in loop_messages -%}
266
+ {#- At this point only assistant/user/tool messages should remain #}
267
+ {%- if message.role == 'assistant' -%}
268
+ {#- Checks to ensure the messages are being passed in the format we expect #}
269
+ {%- if "content" in message %}
270
+ {%- if "<|channel|>analysis<|message|>" in message.content or "<|channel|>final<|message|>" in message.content %}
271
+ {{- raise_exception("You have passed a message containing <|channel|> tags in the content field. Instead of doing this, you should pass analysis messages (the string between '<|message|>' and '<|end|>') in the 'thinking' field, and final messages (the string between '<|message|>' and '<|end|>') in the 'content' field.") }}
272
+ {%- endif %}
273
+ {%- endif %}
274
+ {%- if "thinking" in message %}
275
+ {%- if "<|channel|>analysis<|message|>" in message.thinking or "<|channel|>final<|message|>" in message.thinking %}
276
+ {{- raise_exception("You have passed a message containing <|channel|> tags in the thinking field. Instead of doing this, you should pass analysis messages (the string between '<|message|>' and '<|end|>') in the 'thinking' field, and final messages (the string between '<|message|>' and '<|end|>') in the 'content' field.") }}
277
+ {%- endif %}
278
+ {%- endif %}
279
+ {%- if "tool_calls" in message %}
280
+ {#- We need very careful handling here - we want to drop the tool call analysis message if the model #}
281
+ {#- has output a later <|final|> message, but otherwise we want to retain it. This is the only case #}
282
+ {#- when we render CoT/analysis messages in inference. #}
283
+ {%- set future_final_message = namespace(found=false) %}
284
+ {%- for future_message in loop_messages[loop.index:] %}
285
+ {%- if future_message.role == 'assistant' and "tool_calls" not in future_message %}
286
+ {%- set future_final_message.found = true %}
287
+ {%- endif %}
288
+ {%- endfor %}
289
+ {#- We assume max 1 tool call per message, and so we infer the tool call name #}
290
+ {#- in "tool" messages from the most recent assistant tool call name #}
291
+ {%- set tool_call = message.tool_calls[0] %}
292
+ {%- if tool_call.function %}
293
+ {%- set tool_call = tool_call.function %}
294
+ {%- endif %}
295
+ {%- if message.content and message.thinking %}
296
+ {{- raise_exception("Cannot pass both content and thinking in an assistant message with tool calls! Put the analysis message in one or the other, but not both.") }}
297
+ {%- elif message.content and not future_final_message.found %}
298
+ {{- "<|start|>assistant<|channel|>analysis<|message|>" + message.content + "<|end|>" }}
299
+ {%- elif message.thinking and not future_final_message.found %}
300
+ {{- "<|start|>assistant<|channel|>analysis<|message|>" + message.thinking + "<|end|>" }}
301
+ {%- endif %}
302
+ {{- "<|start|>assistant to=" }}
303
+ {{- "functions." + tool_call.name + "<|channel|>commentary " }}
304
+ {{- (tool_call.content_type if tool_call.content_type is defined else "json") + "<|message|>" }}
305
+ {{- tool_call.arguments|tojson }}
306
+ {{- "<|call|>" }}
307
+ {%- set last_tool_call.name = tool_call.name %}
308
+ {%- elif loop.last and not add_generation_prompt %}
309
+ {#- Only render the CoT if the final turn is an assistant turn and add_generation_prompt is false #}
310
+ {#- This is a situation that should only occur in training, never in inference. #}
311
+ {%- if "thinking" in message %}
312
+ {{- "<|start|>assistant<|channel|>analysis<|message|>" + message.thinking + "<|end|>" }}
313
+ {%- endif %}
314
+ {#- <|return|> indicates the end of generation, but <|end|> does not #}
315
+ {#- <|return|> should never be an input to the model, but we include it as the final token #}
316
+ {#- when training, so the model learns to emit it. #}
317
+ {{- "<|start|>assistant<|channel|>final<|message|>" + message.content + "<|return|>" }}
318
+ {%- else %}
319
+ {#- CoT is dropped during all previous turns, so we never render it for inference #}
320
+ {{- "<|start|>assistant<|channel|>final<|message|>" + message.content + "<|end|>" }}
321
+ {%- set last_tool_call.name = none %}
322
+ {%- endif %}
323
+ {%- elif message.role == 'tool' -%}
324
+ {%- if last_tool_call.name is none %}
325
+ {{- raise_exception("Message has tool role, but there was no previous assistant message with a tool call!") }}
326
+ {%- endif %}
327
+ {{- "<|start|>functions." + last_tool_call.name }}
328
+ {{- " to=assistant<|channel|>commentary<|message|>" + message.content|tojson + "<|end|>" }}
329
+ {%- elif message.role == 'user' -%}
330
+ {{- "<|start|>user<|message|>" + message.content + "<|end|>" }}
331
+ {%- endif -%}
332
+ {%- endfor -%}
333
+
334
+ {#- Generation prompt #}
335
+ {%- if add_generation_prompt -%}
336
+ <|start|>assistant
337
+ {%- endif -%}
checkpoint-3000/config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "dtype": "float32",
9
+ "eos_token_id": 2,
10
+ "head_dim": 128,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 4096,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 14336,
15
+ "max_position_embeddings": 65536,
16
+ "mlp_bias": false,
17
+ "model_type": "llama",
18
+ "num_attention_heads": 32,
19
+ "num_hidden_layers": 32,
20
+ "num_key_value_heads": 8,
21
+ "pad_token_id": 4,
22
+ "pretraining_tp": 1,
23
+ "rms_norm_eps": 1e-06,
24
+ "rope_parameters": {
25
+ "rope_theta": 500000,
26
+ "rope_type": "default"
27
+ },
28
+ "tie_word_embeddings": false,
29
+ "transformers_version": "5.12.1",
30
+ "use_cache": false,
31
+ "vocab_size": 196608
32
+ }
checkpoint-3000/generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": [
5
+ 2
6
+ ],
7
+ "pad_token_id": 4,
8
+ "transformers_version": "5.12.1"
9
+ }
checkpoint-3000/llmjp4_harmony.py ADDED
@@ -0,0 +1,129 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Generic parser for OpenAI Harmony format.
2
+
3
+ from dataclasses import dataclass
4
+ from enum import Enum
5
+ from typing import Iterator, Sequence
6
+
7
+ from transformers import PreTrainedTokenizerBase as TokenizerLike
8
+
9
+
10
+ class HarmonyMessageEndType(Enum):
11
+ INCOMPLETE = 0
12
+ END = 1
13
+ CALL = 2
14
+
15
+
16
+ @dataclass(frozen=True)
17
+ class HarmonySequence:
18
+ """A data class representing a sequence of tokens in the Harmony format."""
19
+ token_ids: list[int]
20
+ start: int # Start position of the sequence in the original token sequence
21
+
22
+
23
+ @dataclass(frozen=True)
24
+ class HarmonyMessage:
25
+ """A data class representing a message in the Harmony format."""
26
+ end: HarmonyMessageEndType
27
+ role: HarmonySequence | None = None
28
+ channel: HarmonySequence | None = None
29
+ constrain: HarmonySequence | None = None
30
+ content: HarmonySequence | None = None
31
+
32
+
33
+ class HarmonyMessageParser:
34
+ """A parser that performs lexical analysis to extract Harmony messages."""
35
+
36
+ def __init__(self, tokenizer: TokenizerLike):
37
+ vocab = tokenizer.get_vocab()
38
+ self._begin_map = {
39
+ vocab["<|start|>"]: "role",
40
+ vocab["<|channel|>"]: "channel",
41
+ vocab["<|constrain|>"]: "constrain",
42
+ vocab["<|message|>"]: "content",
43
+ }
44
+ self._end_map = {
45
+ vocab["<|end|>"]: HarmonyMessageEndType.END,
46
+ vocab["<|return|>"]: HarmonyMessageEndType.END,
47
+ vocab["<|call|>"]: HarmonyMessageEndType.CALL,
48
+ }
49
+
50
+ def iter_messages(self, token_ids: Sequence[int]) -> Iterator[HarmonyMessage]:
51
+ """
52
+ Parse given token ids into messages.
53
+
54
+ Args:
55
+ token_ids: A sequence of token ids to be parsed.
56
+
57
+ Yields:
58
+ Detected HarmonyMessages.
59
+ """
60
+
61
+ message_dict: dict[str, HarmonySequence] = {}
62
+ section: str | None = None # None indicates out-of-message.
63
+ text_ids: list[int] = []
64
+ text_start: int | None = None
65
+
66
+ for token_position, token_id in enumerate(token_ids):
67
+ if token_id in self._begin_map:
68
+ if section is not None:
69
+ message_dict[section] = HarmonySequence(
70
+ token_ids=text_ids,
71
+ start=text_start,
72
+ )
73
+ section = self._begin_map[token_id]
74
+ text_ids = []
75
+ text_start = token_position + 1
76
+
77
+ elif token_id in self._end_map:
78
+ if section is not None:
79
+ message_dict[section] = HarmonySequence(
80
+ token_ids=text_ids,
81
+ start=text_start,
82
+ )
83
+
84
+ yield HarmonyMessage(**message_dict, end=self._end_map[token_id])
85
+
86
+ message_dict = {}
87
+ section = None
88
+ text_ids = []
89
+ text_start = None
90
+
91
+ else:
92
+ if section is not None:
93
+ text_ids.append(token_id)
94
+
95
+ if section is not None:
96
+ message_dict[section] = HarmonySequence(
97
+ token_ids=text_ids,
98
+ start=text_start,
99
+ )
100
+ yield HarmonyMessage(**message_dict, end=HarmonyMessageEndType.INCOMPLETE)
101
+
102
+ def get_all_messages(self, token_ids: Sequence[int]) -> list[HarmonyMessage]:
103
+ """
104
+ Parse given token ids into messages.
105
+
106
+ Args:
107
+ token_ids: A sequence of token ids to be parsed.
108
+
109
+ Returns:
110
+ A list of detected HarmonyMessages.
111
+ """
112
+ return list(self.iter_messages(token_ids))
113
+
114
+ def reverse_iter_messages(self, token_ids: Sequence[int]) -> Iterator[HarmonyMessage]:
115
+ """
116
+ Parse given token ids into messages in reverse order.
117
+
118
+ Args:
119
+ token_ids: A sequence of token ids to be parsed.
120
+
121
+ Yields:
122
+ Detected HarmonyMessages in reverse order.
123
+ """
124
+ end_position = len(token_ids)
125
+
126
+ for i in range(len(token_ids) - 1, -1, -1):
127
+ if token_ids[i] == self._start_id:
128
+ yield next(self.iter_messages(token_ids[i:end_position]))
129
+ end_position = i
checkpoint-3000/llmjp4_tokenizer.py ADDED
@@ -0,0 +1,101 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # llm-jp-4 tokenizer
2
+
3
+ from collections.abc import Sequence
4
+ import os
5
+
6
+ from transformers import LlamaTokenizerFast
7
+ from tokenizers import Tokenizer
8
+
9
+ from .llmjp4_harmony import HarmonyMessageParser, HarmonyMessage
10
+
11
+
12
+ class Llmjp4Tokenizer(LlamaTokenizerFast):
13
+ _HARMONY_TOKENS: set[str] = {
14
+ "<|start|>",
15
+ "<|message|>",
16
+ "<|channel|>",
17
+ "<|constrain|>",
18
+ "<|end|>",
19
+ "<|return|>",
20
+ "<|call|>",
21
+ }
22
+
23
+ # NOTE(odashi):
24
+ # Response schemas are not recognized automatically.
25
+ # We need to define them manually.
26
+ # https://github.com/huggingface/trl/issues/4609
27
+ _RESPONSE_SCHEMA = {
28
+ "type": "object",
29
+ "properties": {
30
+ "role": {"const": "assistant"},
31
+ "content": {"type": "string", "x-regex": r"<\|channel\|>final<\|message\|>(.*?)(?:<\|end\|>|<\|return\|>|$)"},
32
+ "thinking": {"type": "string", "x-regex": r"<\|channel\|>analysis<\|message\|>(.*?)<\|end\|>"},
33
+ "tool_calls": {
34
+ "x-regex-iterator": r"<\|channel\|>commentary (to=functions\..*?<\|message\|>.*?)(?:<\|call\|>|$)",
35
+ "type": "array",
36
+ "items": {
37
+ "type": "object",
38
+ "properties": {
39
+ "type": {"const": "function"},
40
+ "function": {
41
+ "type": "object",
42
+ "properties": {
43
+ "name": {"type": "string", "x-regex": r"^to=functions\.(\w+)"},
44
+ "arguments": {
45
+ "type": "object",
46
+ "x-regex": r"<\|message\|>(.*)",
47
+ "x-parser": "json",
48
+ "additionalProperties": {"type": "any"},
49
+ },
50
+ },
51
+ },
52
+ },
53
+ },
54
+ },
55
+ },
56
+ }
57
+
58
+ @classmethod
59
+ def convert_to_native_format(cls, **kwargs):
60
+ # NOTE(odashi):
61
+ # Workaround for transformers 5.x.
62
+ # Guaranteeing the same inner behavior with TokenizersBackend.
63
+ # https://github.com/huggingface/transformers/blob/7d9754a05193eb79b1d86aa744b622b8068008cd/src/transformers/tokenization_utils_tokenizers.py#L110-L116
64
+ local_kwargs = dict(kwargs)
65
+ fast_tokenizer_file = local_kwargs.pop("tokenizer_file", None)
66
+ if fast_tokenizer_file is None or not os.path.isfile(fast_tokenizer_file):
67
+ raise ValueError("Tokenizer file must exist.")
68
+
69
+ local_kwargs["tokenizer_object"] = Tokenizer.from_file(fast_tokenizer_file)
70
+ return local_kwargs
71
+
72
+ def __init__(self, *args, **kwargs):
73
+ super().__init__(*args, **kwargs)
74
+
75
+ self.response_schema = self._RESPONSE_SCHEMA
76
+
77
+ self._harmony_token_ids = {
78
+ self.convert_tokens_to_ids(token)
79
+ for token in self._HARMONY_TOKENS
80
+ }
81
+
82
+ def _decode(self, token_ids: int | list[int], *args, **kwargs):
83
+ if isinstance(token_ids, int):
84
+ token_ids = [token_ids]
85
+
86
+ result: list[str] = []
87
+ prev_pos = 0
88
+
89
+ # NOTE(odashi):
90
+ # Ensure that text tokens are decoded without preceding Harmony tokens
91
+ # to avoid incorrect addition of whitespaces.
92
+ for pos, token_id in enumerate(token_ids, start=1):
93
+ if token_id in self._harmony_token_ids or pos == len(token_ids):
94
+ result.append(super()._decode(token_ids[prev_pos:pos], *args, **kwargs))
95
+ prev_pos = pos
96
+
97
+ return "".join(result)
98
+
99
+ def parse_harmony_message(self, token_ids: Sequence[int]) -> list[HarmonyMessage]:
100
+ """Helper function to parse token IDs into Harmony messages."""
101
+ return HarmonyMessageParser(self).get_all_messages(token_ids)
checkpoint-3000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c4a1acba5c71256e2290e997c99ea8e820a4c06a26e8dcb3931c58dc031f294c
3
+ size 34360837408
checkpoint-3000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c4b27ab4cd1410523fa93f23504b69281ff25cf7f4756638a439f7f95ba1c68f
3
+ size 16389
checkpoint-3000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b4c51e1faf27a7579852207779a2a2419e29f10086fb03daf2a071d76f9aa66
3
+ size 16389
checkpoint-3000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:461c4f18c4a9e5ea1306623fd279ecbe34fb58d34e39f12345703d5dc4f6ae16
3
+ size 16389
checkpoint-3000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ea60285bf794d12b45bdbbc17d62f1fdd2c7b31f8d841de76b3b726d3ff686d3
3
+ size 16389
checkpoint-3000/rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2e35040709f0585bd45e7de22541a68f700a7eb5bb4fc1988fb2446dd2e4fbcf
3
+ size 16389
checkpoint-3000/rng_state_5.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:71a4ebd3cccae9386eaf42d9b8f60e96c894d35abaeaf75aa34d6386af344f0d
3
+ size 16389
checkpoint-3000/rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e990d2633dbecb8ade2cff16be5fc46e8acc5f0a78525dbdcd1c4a42dc93186f
3
+ size 16389
checkpoint-3000/rng_state_7.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:abe8282828a39414c7542e9ccebecf61132009dd563028657ddd6d169d52b3c7
3
+ size 16389
checkpoint-3000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7639e00276b3bb700396ddbcac527b74d64520d85e8d1d2afe314ef24f1ca451
3
+ size 1465
checkpoint-3000/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15d5f21ae725fc96a5766271720bef30e132bfba89c11820a1e8612e5f643426
3
+ size 12868031
checkpoint-3000/tokenizer_config.json ADDED
@@ -0,0 +1,74 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": null,
3
+ "auto_map": {
4
+ "AutoTokenizer": [
5
+ "llmjp4_tokenizer.Llmjp4Tokenizer",
6
+ null
7
+ ]
8
+ },
9
+ "backend": "tokenizers",
10
+ "bos_token": "<|startoftext|>",
11
+ "clean_up_tokenization_spaces": false,
12
+ "cls_token": "<|cls|>",
13
+ "eod_token": "<|eod|>",
14
+ "eos_token": "<|return|>",
15
+ "extra_ids": 0,
16
+ "is_local": false,
17
+ "local_files_only": false,
18
+ "mask_token": "<|mask|>",
19
+ "model_max_length": 1000000000000000019884624838656,
20
+ "model_specific_special_tokens": {
21
+ "eod_token": "<|eod|>"
22
+ },
23
+ "pad_token": "<|endoftext|>",
24
+ "response_schema": {
25
+ "properties": {
26
+ "content": {
27
+ "type": "string",
28
+ "x-regex": "<\\|channel\\|>final<\\|message\\|>(.*?)(?:<\\|end\\|>|<\\|return\\|>|$)"
29
+ },
30
+ "role": {
31
+ "const": "assistant"
32
+ },
33
+ "thinking": {
34
+ "type": "string",
35
+ "x-regex": "<\\|channel\\|>analysis<\\|message\\|>(.*?)<\\|end\\|>"
36
+ },
37
+ "tool_calls": {
38
+ "items": {
39
+ "properties": {
40
+ "function": {
41
+ "properties": {
42
+ "arguments": {
43
+ "additionalProperties": {
44
+ "type": "any"
45
+ },
46
+ "type": "object",
47
+ "x-parser": "json",
48
+ "x-regex": "<\\|message\\|>(.*)"
49
+ },
50
+ "name": {
51
+ "type": "string",
52
+ "x-regex": "^to=functions\\.(\\w+)"
53
+ }
54
+ },
55
+ "type": "object"
56
+ },
57
+ "type": {
58
+ "const": "function"
59
+ }
60
+ },
61
+ "type": "object"
62
+ },
63
+ "type": "array",
64
+ "x-regex-iterator": "<\\|channel\\|>commentary (to=functions\\..*?<\\|message\\|>.*?)(?:<\\|call\\|>|$)"
65
+ }
66
+ },
67
+ "type": "object"
68
+ },
69
+ "sep_token": "<|sep|>",
70
+ "sp_model_kwargs": {},
71
+ "tokenizer_class": "Llmjp4Tokenizer",
72
+ "unk_token": "<|unk|>",
73
+ "use_default_system_prompt": false
74
+ }
checkpoint-3000/trainer_state.json ADDED
@@ -0,0 +1,2134 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.07031576171746248,
6
+ "eval_steps": 500,
7
+ "global_step": 3000,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0002343858723915416,
14
+ "grad_norm": 7.478665351867676,
15
+ "learning_rate": 1.40625e-07,
16
+ "loss": 0.6732528686523438,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.0004687717447830832,
21
+ "grad_norm": 7.584623336791992,
22
+ "learning_rate": 2.96875e-07,
23
+ "loss": 0.66597900390625,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.0007031576171746248,
28
+ "grad_norm": 6.306713581085205,
29
+ "learning_rate": 4.53125e-07,
30
+ "loss": 0.6197677612304687,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.0009375434895661664,
35
+ "grad_norm": 4.607383728027344,
36
+ "learning_rate": 6.09375e-07,
37
+ "loss": 0.5319419860839844,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.001171929361957708,
42
+ "grad_norm": 2.3062853813171387,
43
+ "learning_rate": 7.656250000000001e-07,
44
+ "loss": 0.42061538696289064,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.0014063152343492496,
49
+ "grad_norm": 1.4863656759262085,
50
+ "learning_rate": 9.218750000000002e-07,
51
+ "loss": 0.3130531311035156,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.0016407011067407912,
56
+ "grad_norm": 0.8237831592559814,
57
+ "learning_rate": 1.0781250000000002e-06,
58
+ "loss": 0.23311138153076172,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.0018750869791323327,
63
+ "grad_norm": 0.6643654704093933,
64
+ "learning_rate": 1.2343750000000001e-06,
65
+ "loss": 0.2020857810974121,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.0021094728515238742,
70
+ "grad_norm": 0.5149619579315186,
71
+ "learning_rate": 1.3906250000000001e-06,
72
+ "loss": 0.18220624923706055,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.002343858723915416,
77
+ "grad_norm": 0.5113442540168762,
78
+ "learning_rate": 1.5468750000000001e-06,
79
+ "loss": 0.16905088424682618,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.0025782445963069577,
84
+ "grad_norm": 0.5090538859367371,
85
+ "learning_rate": 1.703125e-06,
86
+ "loss": 0.16316838264465333,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.0028126304686984993,
91
+ "grad_norm": 0.4110700190067291,
92
+ "learning_rate": 1.8593750000000003e-06,
93
+ "loss": 0.1580258846282959,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.003047016341090041,
98
+ "grad_norm": 0.5005377531051636,
99
+ "learning_rate": 2.0156250000000003e-06,
100
+ "loss": 0.14841842651367188,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.0032814022134815823,
105
+ "grad_norm": 0.4914725124835968,
106
+ "learning_rate": 2.1718750000000003e-06,
107
+ "loss": 0.14541361331939698,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.003515788085873124,
112
+ "grad_norm": 0.5155150294303894,
113
+ "learning_rate": 2.3281250000000003e-06,
114
+ "loss": 0.14492216110229492,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.0037501739582646654,
119
+ "grad_norm": 0.4690794050693512,
120
+ "learning_rate": 2.4843750000000002e-06,
121
+ "loss": 0.14195268154144286,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.003984559830656207,
126
+ "grad_norm": 0.50107342004776,
127
+ "learning_rate": 2.640625e-06,
128
+ "loss": 0.1428708553314209,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.0042189457030477485,
133
+ "grad_norm": 0.48859548568725586,
134
+ "learning_rate": 2.796875e-06,
135
+ "loss": 0.14226298332214354,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.0044533315754392904,
140
+ "grad_norm": 0.5123774409294128,
141
+ "learning_rate": 2.953125e-06,
142
+ "loss": 0.1370567798614502,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.004687717447830832,
147
+ "grad_norm": 0.4833315312862396,
148
+ "learning_rate": 3.109375e-06,
149
+ "loss": 0.1390127420425415,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.0049221033202223735,
154
+ "grad_norm": 0.43915465474128723,
155
+ "learning_rate": 3.265625e-06,
156
+ "loss": 0.13475892543792725,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.0051564891926139155,
161
+ "grad_norm": 0.46104806661605835,
162
+ "learning_rate": 3.421875e-06,
163
+ "loss": 0.13139626979827881,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.005390875065005457,
168
+ "grad_norm": 0.5122255086898804,
169
+ "learning_rate": 3.578125e-06,
170
+ "loss": 0.13110661506652832,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.0056252609373969985,
175
+ "grad_norm": 0.48517927527427673,
176
+ "learning_rate": 3.734375e-06,
177
+ "loss": 0.12811717987060547,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.00585964680978854,
182
+ "grad_norm": 0.4835013449192047,
183
+ "learning_rate": 3.890625e-06,
184
+ "loss": 0.13097090721130372,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.006094032682180082,
189
+ "grad_norm": 0.4820455312728882,
190
+ "learning_rate": 4.046875e-06,
191
+ "loss": 0.12818758487701415,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.006328418554571624,
196
+ "grad_norm": 0.44813403487205505,
197
+ "learning_rate": 4.2031250000000005e-06,
198
+ "loss": 0.12801458835601806,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.006562804426963165,
203
+ "grad_norm": 0.5279086232185364,
204
+ "learning_rate": 4.359375e-06,
205
+ "loss": 0.12751049995422364,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.006797190299354707,
210
+ "grad_norm": 0.42281895875930786,
211
+ "learning_rate": 4.5156250000000005e-06,
212
+ "loss": 0.13442916870117189,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.007031576171746248,
217
+ "grad_norm": 0.4167884588241577,
218
+ "learning_rate": 4.671875e-06,
219
+ "loss": 0.12415478229522706,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.00726596204413779,
224
+ "grad_norm": 0.4328134059906006,
225
+ "learning_rate": 4.8281250000000005e-06,
226
+ "loss": 0.12891750335693358,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.007500347916529331,
231
+ "grad_norm": 0.45806413888931274,
232
+ "learning_rate": 4.984375e-06,
233
+ "loss": 0.12102985382080078,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.007734733788920873,
238
+ "grad_norm": 0.5078896880149841,
239
+ "learning_rate": 5.1406250000000004e-06,
240
+ "loss": 0.12243216037750244,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.007969119661312415,
245
+ "grad_norm": 0.527930498123169,
246
+ "learning_rate": 5.296875e-06,
247
+ "loss": 0.12513620853424073,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.008203505533703956,
252
+ "grad_norm": 0.431318461894989,
253
+ "learning_rate": 5.453125e-06,
254
+ "loss": 0.12151198387145996,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.008437891406095497,
259
+ "grad_norm": 0.4996262788772583,
260
+ "learning_rate": 5.609375e-06,
261
+ "loss": 0.11847388744354248,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.00867227727848704,
266
+ "grad_norm": 0.4818030893802643,
267
+ "learning_rate": 5.765625e-06,
268
+ "loss": 0.1227838158607483,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.008906663150878581,
273
+ "grad_norm": 0.4577305018901825,
274
+ "learning_rate": 5.921875e-06,
275
+ "loss": 0.1199771523475647,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.009141049023270122,
280
+ "grad_norm": 0.48360294103622437,
281
+ "learning_rate": 6.078125e-06,
282
+ "loss": 0.11903550624847412,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.009375434895661665,
287
+ "grad_norm": 0.47577178478240967,
288
+ "learning_rate": 6.234375e-06,
289
+ "loss": 0.12074012756347656,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.009609820768053206,
294
+ "grad_norm": 0.4833427369594574,
295
+ "learning_rate": 6.390625e-06,
296
+ "loss": 0.12326394319534302,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.009844206640444747,
301
+ "grad_norm": 0.43909400701522827,
302
+ "learning_rate": 6.546875e-06,
303
+ "loss": 0.11921967267990112,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.010078592512836288,
308
+ "grad_norm": 0.5028750896453857,
309
+ "learning_rate": 6.703125e-06,
310
+ "loss": 0.12188574075698852,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.010312978385227831,
315
+ "grad_norm": 0.47389668226242065,
316
+ "learning_rate": 6.859375000000001e-06,
317
+ "loss": 0.1192385196685791,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.010547364257619372,
322
+ "grad_norm": 0.5239953398704529,
323
+ "learning_rate": 7.015625e-06,
324
+ "loss": 0.11499193906784058,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.010781750130010913,
329
+ "grad_norm": 0.5061202645301819,
330
+ "learning_rate": 7.171875000000001e-06,
331
+ "loss": 0.11584588289260864,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.011016136002402456,
336
+ "grad_norm": 0.4635751247406006,
337
+ "learning_rate": 7.328125e-06,
338
+ "loss": 0.11434779167175294,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.011250521874793997,
343
+ "grad_norm": 0.4353443682193756,
344
+ "learning_rate": 7.484375000000001e-06,
345
+ "loss": 0.11755821704864503,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.011484907747185538,
350
+ "grad_norm": 0.41028693318367004,
351
+ "learning_rate": 7.640625000000001e-06,
352
+ "loss": 0.11704204082489014,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.01171929361957708,
357
+ "grad_norm": 0.4426499009132385,
358
+ "learning_rate": 7.796875e-06,
359
+ "loss": 0.11789888143539429,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.011953679491968622,
364
+ "grad_norm": 0.45650583505630493,
365
+ "learning_rate": 7.953125e-06,
366
+ "loss": 0.11612817049026489,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.012188065364360163,
371
+ "grad_norm": 0.48528221249580383,
372
+ "learning_rate": 8.109375e-06,
373
+ "loss": 0.11835837364196777,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.012422451236751704,
378
+ "grad_norm": 0.49201133847236633,
379
+ "learning_rate": 8.265625000000001e-06,
380
+ "loss": 0.11796109676361084,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.012656837109143247,
385
+ "grad_norm": 0.41120538115501404,
386
+ "learning_rate": 8.421875e-06,
387
+ "loss": 0.11465147733688355,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 0.012891222981534788,
392
+ "grad_norm": 0.42535409331321716,
393
+ "learning_rate": 8.578125e-06,
394
+ "loss": 0.11844432353973389,
395
+ "step": 550
396
+ },
397
+ {
398
+ "epoch": 0.01312560885392633,
399
+ "grad_norm": 0.44362396001815796,
400
+ "learning_rate": 8.734375e-06,
401
+ "loss": 0.11715601682662964,
402
+ "step": 560
403
+ },
404
+ {
405
+ "epoch": 0.01335999472631787,
406
+ "grad_norm": 0.441021203994751,
407
+ "learning_rate": 8.890625000000001e-06,
408
+ "loss": 0.11989809274673462,
409
+ "step": 570
410
+ },
411
+ {
412
+ "epoch": 0.013594380598709413,
413
+ "grad_norm": 0.4338180720806122,
414
+ "learning_rate": 9.046875e-06,
415
+ "loss": 0.11536179780960083,
416
+ "step": 580
417
+ },
418
+ {
419
+ "epoch": 0.013828766471100954,
420
+ "grad_norm": 0.43819499015808105,
421
+ "learning_rate": 9.203125e-06,
422
+ "loss": 0.11345911026000977,
423
+ "step": 590
424
+ },
425
+ {
426
+ "epoch": 0.014063152343492496,
427
+ "grad_norm": 0.4360722601413727,
428
+ "learning_rate": 9.359375e-06,
429
+ "loss": 0.11602315902709961,
430
+ "step": 600
431
+ },
432
+ {
433
+ "epoch": 0.014297538215884038,
434
+ "grad_norm": 0.4287785589694977,
435
+ "learning_rate": 9.515625000000001e-06,
436
+ "loss": 0.11890817880630493,
437
+ "step": 610
438
+ },
439
+ {
440
+ "epoch": 0.01453192408827558,
441
+ "grad_norm": 0.4322708547115326,
442
+ "learning_rate": 9.671875000000001e-06,
443
+ "loss": 0.11710528135299683,
444
+ "step": 620
445
+ },
446
+ {
447
+ "epoch": 0.01476630996066712,
448
+ "grad_norm": 0.4133163392543793,
449
+ "learning_rate": 9.828125e-06,
450
+ "loss": 0.11247456073760986,
451
+ "step": 630
452
+ },
453
+ {
454
+ "epoch": 0.015000695833058662,
455
+ "grad_norm": 0.4161861538887024,
456
+ "learning_rate": 9.984375e-06,
457
+ "loss": 0.1176948070526123,
458
+ "step": 640
459
+ },
460
+ {
461
+ "epoch": 0.015235081705450204,
462
+ "grad_norm": 0.4499410092830658,
463
+ "learning_rate": 1.0140625000000003e-05,
464
+ "loss": 0.11162110567092895,
465
+ "step": 650
466
+ },
467
+ {
468
+ "epoch": 0.015469467577841746,
469
+ "grad_norm": 0.4393712878227234,
470
+ "learning_rate": 1.0296875000000001e-05,
471
+ "loss": 0.11646484136581421,
472
+ "step": 660
473
+ },
474
+ {
475
+ "epoch": 0.01570385345023329,
476
+ "grad_norm": 0.43581581115722656,
477
+ "learning_rate": 1.0453125000000002e-05,
478
+ "loss": 0.11804448366165161,
479
+ "step": 670
480
+ },
481
+ {
482
+ "epoch": 0.01593823932262483,
483
+ "grad_norm": 0.5233722925186157,
484
+ "learning_rate": 1.0609375000000002e-05,
485
+ "loss": 0.11381900310516357,
486
+ "step": 680
487
+ },
488
+ {
489
+ "epoch": 0.01617262519501637,
490
+ "grad_norm": 0.4110233187675476,
491
+ "learning_rate": 1.0765625000000002e-05,
492
+ "loss": 0.11446200609207154,
493
+ "step": 690
494
+ },
495
+ {
496
+ "epoch": 0.01640701106740791,
497
+ "grad_norm": 0.399118036031723,
498
+ "learning_rate": 1.0921875000000001e-05,
499
+ "loss": 0.11271647214889527,
500
+ "step": 700
501
+ },
502
+ {
503
+ "epoch": 0.016641396939799453,
504
+ "grad_norm": 0.46171241998672485,
505
+ "learning_rate": 1.1078125000000002e-05,
506
+ "loss": 0.11184619665145874,
507
+ "step": 710
508
+ },
509
+ {
510
+ "epoch": 0.016875782812190994,
511
+ "grad_norm": 0.45813125371932983,
512
+ "learning_rate": 1.1234375000000002e-05,
513
+ "loss": 0.11524376869201661,
514
+ "step": 720
515
+ },
516
+ {
517
+ "epoch": 0.01711016868458254,
518
+ "grad_norm": 0.3914071023464203,
519
+ "learning_rate": 1.1390625000000002e-05,
520
+ "loss": 0.11292357444763183,
521
+ "step": 730
522
+ },
523
+ {
524
+ "epoch": 0.01734455455697408,
525
+ "grad_norm": 0.3806169033050537,
526
+ "learning_rate": 1.1546875000000001e-05,
527
+ "loss": 0.11865770816802979,
528
+ "step": 740
529
+ },
530
+ {
531
+ "epoch": 0.01757894042936562,
532
+ "grad_norm": 0.6183323264122009,
533
+ "learning_rate": 1.1703125000000002e-05,
534
+ "loss": 0.1167568564414978,
535
+ "step": 750
536
+ },
537
+ {
538
+ "epoch": 0.017813326301757162,
539
+ "grad_norm": 0.39449653029441833,
540
+ "learning_rate": 1.1859375000000002e-05,
541
+ "loss": 0.11325149536132813,
542
+ "step": 760
543
+ },
544
+ {
545
+ "epoch": 0.018047712174148703,
546
+ "grad_norm": 0.37599578499794006,
547
+ "learning_rate": 1.2015625000000002e-05,
548
+ "loss": 0.11163055896759033,
549
+ "step": 770
550
+ },
551
+ {
552
+ "epoch": 0.018282098046540244,
553
+ "grad_norm": 0.38267573714256287,
554
+ "learning_rate": 1.2171875000000001e-05,
555
+ "loss": 0.11489014625549317,
556
+ "step": 780
557
+ },
558
+ {
559
+ "epoch": 0.018516483918931785,
560
+ "grad_norm": 0.37311288714408875,
561
+ "learning_rate": 1.2328125000000002e-05,
562
+ "loss": 0.11194641590118408,
563
+ "step": 790
564
+ },
565
+ {
566
+ "epoch": 0.01875086979132333,
567
+ "grad_norm": 0.411432683467865,
568
+ "learning_rate": 1.2484375000000002e-05,
569
+ "loss": 0.11317495107650757,
570
+ "step": 800
571
+ },
572
+ {
573
+ "epoch": 0.01898525566371487,
574
+ "grad_norm": 0.43839067220687866,
575
+ "learning_rate": 1.2640625000000002e-05,
576
+ "loss": 0.10759581327438354,
577
+ "step": 810
578
+ },
579
+ {
580
+ "epoch": 0.019219641536106412,
581
+ "grad_norm": 0.3859034478664398,
582
+ "learning_rate": 1.2796875000000003e-05,
583
+ "loss": 0.11314148902893066,
584
+ "step": 820
585
+ },
586
+ {
587
+ "epoch": 0.019454027408497953,
588
+ "grad_norm": 0.4093567728996277,
589
+ "learning_rate": 1.2953125000000001e-05,
590
+ "loss": 0.11214956045150756,
591
+ "step": 830
592
+ },
593
+ {
594
+ "epoch": 0.019688413280889494,
595
+ "grad_norm": 0.4202522337436676,
596
+ "learning_rate": 1.3109375000000002e-05,
597
+ "loss": 0.10863385200500489,
598
+ "step": 840
599
+ },
600
+ {
601
+ "epoch": 0.019922799153281035,
602
+ "grad_norm": 0.37031644582748413,
603
+ "learning_rate": 1.3265625000000002e-05,
604
+ "loss": 0.11562912464141846,
605
+ "step": 850
606
+ },
607
+ {
608
+ "epoch": 0.020157185025672576,
609
+ "grad_norm": 0.372043251991272,
610
+ "learning_rate": 1.3421875000000003e-05,
611
+ "loss": 0.11000908613204956,
612
+ "step": 860
613
+ },
614
+ {
615
+ "epoch": 0.02039157089806412,
616
+ "grad_norm": 0.39329200983047485,
617
+ "learning_rate": 1.3578125000000001e-05,
618
+ "loss": 0.11322653293609619,
619
+ "step": 870
620
+ },
621
+ {
622
+ "epoch": 0.020625956770455662,
623
+ "grad_norm": 0.3731657564640045,
624
+ "learning_rate": 1.3734375000000002e-05,
625
+ "loss": 0.11117322444915771,
626
+ "step": 880
627
+ },
628
+ {
629
+ "epoch": 0.020860342642847203,
630
+ "grad_norm": 0.3705548644065857,
631
+ "learning_rate": 1.3890625000000002e-05,
632
+ "loss": 0.10953671932220459,
633
+ "step": 890
634
+ },
635
+ {
636
+ "epoch": 0.021094728515238744,
637
+ "grad_norm": 0.38085901737213135,
638
+ "learning_rate": 1.4046875000000003e-05,
639
+ "loss": 0.11695330142974854,
640
+ "step": 900
641
+ },
642
+ {
643
+ "epoch": 0.021329114387630285,
644
+ "grad_norm": 0.4012692868709564,
645
+ "learning_rate": 1.4203125000000001e-05,
646
+ "loss": 0.11438255310058594,
647
+ "step": 910
648
+ },
649
+ {
650
+ "epoch": 0.021563500260021826,
651
+ "grad_norm": 0.3757342994213104,
652
+ "learning_rate": 1.4359375000000002e-05,
653
+ "loss": 0.11287130117416382,
654
+ "step": 920
655
+ },
656
+ {
657
+ "epoch": 0.021797886132413367,
658
+ "grad_norm": 0.38042888045310974,
659
+ "learning_rate": 1.4515625000000002e-05,
660
+ "loss": 0.11060645580291747,
661
+ "step": 930
662
+ },
663
+ {
664
+ "epoch": 0.022032272004804912,
665
+ "grad_norm": 0.3464396297931671,
666
+ "learning_rate": 1.4671875000000003e-05,
667
+ "loss": 0.11261942386627197,
668
+ "step": 940
669
+ },
670
+ {
671
+ "epoch": 0.022266657877196453,
672
+ "grad_norm": 0.3607310652732849,
673
+ "learning_rate": 1.4828125000000001e-05,
674
+ "loss": 0.11483677625656127,
675
+ "step": 950
676
+ },
677
+ {
678
+ "epoch": 0.022501043749587994,
679
+ "grad_norm": 0.37492862343788147,
680
+ "learning_rate": 1.4984375000000002e-05,
681
+ "loss": 0.11412110328674316,
682
+ "step": 960
683
+ },
684
+ {
685
+ "epoch": 0.022735429621979535,
686
+ "grad_norm": 0.36724552512168884,
687
+ "learning_rate": 1.5140625000000002e-05,
688
+ "loss": 0.11400502920150757,
689
+ "step": 970
690
+ },
691
+ {
692
+ "epoch": 0.022969815494371076,
693
+ "grad_norm": 0.3513246774673462,
694
+ "learning_rate": 1.5296875e-05,
695
+ "loss": 0.11101123094558715,
696
+ "step": 980
697
+ },
698
+ {
699
+ "epoch": 0.023204201366762617,
700
+ "grad_norm": 0.39282527565956116,
701
+ "learning_rate": 1.5453125e-05,
702
+ "loss": 0.11446715593338012,
703
+ "step": 990
704
+ },
705
+ {
706
+ "epoch": 0.02343858723915416,
707
+ "grad_norm": 0.3693297505378723,
708
+ "learning_rate": 1.5609375e-05,
709
+ "loss": 0.11360613107681275,
710
+ "step": 1000
711
+ },
712
+ {
713
+ "epoch": 0.023672973111545703,
714
+ "grad_norm": 0.3805910050868988,
715
+ "learning_rate": 1.5765625000000002e-05,
716
+ "loss": 0.1102859616279602,
717
+ "step": 1010
718
+ },
719
+ {
720
+ "epoch": 0.023907358983937244,
721
+ "grad_norm": 0.3862535059452057,
722
+ "learning_rate": 1.5921875000000002e-05,
723
+ "loss": 0.11268872022628784,
724
+ "step": 1020
725
+ },
726
+ {
727
+ "epoch": 0.024141744856328785,
728
+ "grad_norm": 0.404893696308136,
729
+ "learning_rate": 1.6078125000000003e-05,
730
+ "loss": 0.11359611749649048,
731
+ "step": 1030
732
+ },
733
+ {
734
+ "epoch": 0.024376130728720326,
735
+ "grad_norm": 0.38326576352119446,
736
+ "learning_rate": 1.6234375000000003e-05,
737
+ "loss": 0.11227505207061768,
738
+ "step": 1040
739
+ },
740
+ {
741
+ "epoch": 0.024610516601111868,
742
+ "grad_norm": 0.36912357807159424,
743
+ "learning_rate": 1.6390625000000004e-05,
744
+ "loss": 0.11599531173706054,
745
+ "step": 1050
746
+ },
747
+ {
748
+ "epoch": 0.02484490247350341,
749
+ "grad_norm": 0.3696171045303345,
750
+ "learning_rate": 1.6546875e-05,
751
+ "loss": 0.11023976802825927,
752
+ "step": 1060
753
+ },
754
+ {
755
+ "epoch": 0.02507928834589495,
756
+ "grad_norm": 0.3755769729614258,
757
+ "learning_rate": 1.6703125e-05,
758
+ "loss": 0.11145970821380616,
759
+ "step": 1070
760
+ },
761
+ {
762
+ "epoch": 0.025313674218286494,
763
+ "grad_norm": 0.3687419593334198,
764
+ "learning_rate": 1.6859375e-05,
765
+ "loss": 0.11287758350372315,
766
+ "step": 1080
767
+ },
768
+ {
769
+ "epoch": 0.025548060090678035,
770
+ "grad_norm": 0.33784812688827515,
771
+ "learning_rate": 1.7015625000000002e-05,
772
+ "loss": 0.10908894538879395,
773
+ "step": 1090
774
+ },
775
+ {
776
+ "epoch": 0.025782445963069577,
777
+ "grad_norm": 0.3508151173591614,
778
+ "learning_rate": 1.7171875000000002e-05,
779
+ "loss": 0.11538139581680298,
780
+ "step": 1100
781
+ },
782
+ {
783
+ "epoch": 0.026016831835461118,
784
+ "grad_norm": 0.36725053191185,
785
+ "learning_rate": 1.7328125000000003e-05,
786
+ "loss": 0.11482850313186646,
787
+ "step": 1110
788
+ },
789
+ {
790
+ "epoch": 0.02625121770785266,
791
+ "grad_norm": 0.34291842579841614,
792
+ "learning_rate": 1.7484375000000003e-05,
793
+ "loss": 0.11298364400863647,
794
+ "step": 1120
795
+ },
796
+ {
797
+ "epoch": 0.0264856035802442,
798
+ "grad_norm": 0.36184895038604736,
799
+ "learning_rate": 1.7640625000000004e-05,
800
+ "loss": 0.1066980242729187,
801
+ "step": 1130
802
+ },
803
+ {
804
+ "epoch": 0.02671998945263574,
805
+ "grad_norm": 0.4020235240459442,
806
+ "learning_rate": 1.7796875000000004e-05,
807
+ "loss": 0.11315239667892456,
808
+ "step": 1140
809
+ },
810
+ {
811
+ "epoch": 0.026954375325027286,
812
+ "grad_norm": 0.34352606534957886,
813
+ "learning_rate": 1.7953125e-05,
814
+ "loss": 0.10964906215667725,
815
+ "step": 1150
816
+ },
817
+ {
818
+ "epoch": 0.027188761197418827,
819
+ "grad_norm": 0.4462008476257324,
820
+ "learning_rate": 1.8109375e-05,
821
+ "loss": 0.11428066492080688,
822
+ "step": 1160
823
+ },
824
+ {
825
+ "epoch": 0.027423147069810368,
826
+ "grad_norm": 0.3392775058746338,
827
+ "learning_rate": 1.8265625000000002e-05,
828
+ "loss": 0.11005079746246338,
829
+ "step": 1170
830
+ },
831
+ {
832
+ "epoch": 0.02765753294220191,
833
+ "grad_norm": 0.3385438621044159,
834
+ "learning_rate": 1.8421875000000002e-05,
835
+ "loss": 0.11102509498596191,
836
+ "step": 1180
837
+ },
838
+ {
839
+ "epoch": 0.02789191881459345,
840
+ "grad_norm": 0.3650127649307251,
841
+ "learning_rate": 1.8578125000000003e-05,
842
+ "loss": 0.11523213386535644,
843
+ "step": 1190
844
+ },
845
+ {
846
+ "epoch": 0.02812630468698499,
847
+ "grad_norm": 0.3460846245288849,
848
+ "learning_rate": 1.8734375000000003e-05,
849
+ "loss": 0.11034642457962036,
850
+ "step": 1200
851
+ },
852
+ {
853
+ "epoch": 0.028360690559376532,
854
+ "grad_norm": 0.326349675655365,
855
+ "learning_rate": 1.8890625000000003e-05,
856
+ "loss": 0.10953868627548217,
857
+ "step": 1210
858
+ },
859
+ {
860
+ "epoch": 0.028595076431768077,
861
+ "grad_norm": 0.34334462881088257,
862
+ "learning_rate": 1.9046875000000004e-05,
863
+ "loss": 0.11296612024307251,
864
+ "step": 1220
865
+ },
866
+ {
867
+ "epoch": 0.028829462304159618,
868
+ "grad_norm": 0.33681803941726685,
869
+ "learning_rate": 1.9203125e-05,
870
+ "loss": 0.11056774854660034,
871
+ "step": 1230
872
+ },
873
+ {
874
+ "epoch": 0.02906384817655116,
875
+ "grad_norm": 0.3528500199317932,
876
+ "learning_rate": 1.9359375e-05,
877
+ "loss": 0.10871942043304443,
878
+ "step": 1240
879
+ },
880
+ {
881
+ "epoch": 0.0292982340489427,
882
+ "grad_norm": 0.3156096339225769,
883
+ "learning_rate": 1.9515625000000002e-05,
884
+ "loss": 0.1137766718864441,
885
+ "step": 1250
886
+ },
887
+ {
888
+ "epoch": 0.02953261992133424,
889
+ "grad_norm": 0.3579378128051758,
890
+ "learning_rate": 1.9671875000000002e-05,
891
+ "loss": 0.11373578310012818,
892
+ "step": 1260
893
+ },
894
+ {
895
+ "epoch": 0.029767005793725782,
896
+ "grad_norm": 0.315738707780838,
897
+ "learning_rate": 1.9828125000000003e-05,
898
+ "loss": 0.11297824382781982,
899
+ "step": 1270
900
+ },
901
+ {
902
+ "epoch": 0.030001391666117323,
903
+ "grad_norm": 0.3177584111690521,
904
+ "learning_rate": 1.9984375000000003e-05,
905
+ "loss": 0.1135290265083313,
906
+ "step": 1280
907
+ },
908
+ {
909
+ "epoch": 0.030235777538508868,
910
+ "grad_norm": 0.34855231642723083,
911
+ "learning_rate": 1.9999997666172074e-05,
912
+ "loss": 0.11188427209854127,
913
+ "step": 1290
914
+ },
915
+ {
916
+ "epoch": 0.03047016341090041,
917
+ "grad_norm": 0.3580094575881958,
918
+ "learning_rate": 1.9999989598620137e-05,
919
+ "loss": 0.10642122030258179,
920
+ "step": 1300
921
+ },
922
+ {
923
+ "epoch": 0.03070454928329195,
924
+ "grad_norm": 0.37562474608421326,
925
+ "learning_rate": 1.9999975768536158e-05,
926
+ "loss": 0.11398833990097046,
927
+ "step": 1310
928
+ },
929
+ {
930
+ "epoch": 0.03093893515568349,
931
+ "grad_norm": 0.34030941128730774,
932
+ "learning_rate": 1.9999956175928097e-05,
933
+ "loss": 0.10998734235763549,
934
+ "step": 1320
935
+ },
936
+ {
937
+ "epoch": 0.031173321028075032,
938
+ "grad_norm": 0.3101728856563568,
939
+ "learning_rate": 1.9999930820807245e-05,
940
+ "loss": 0.11209770441055297,
941
+ "step": 1330
942
+ },
943
+ {
944
+ "epoch": 0.03140770690046658,
945
+ "grad_norm": 0.34433069825172424,
946
+ "learning_rate": 1.999989970318822e-05,
947
+ "loss": 0.1109668493270874,
948
+ "step": 1340
949
+ },
950
+ {
951
+ "epoch": 0.031642092772858114,
952
+ "grad_norm": 0.32215601205825806,
953
+ "learning_rate": 1.9999862823088946e-05,
954
+ "loss": 0.10847405195236207,
955
+ "step": 1350
956
+ },
957
+ {
958
+ "epoch": 0.03187647864524966,
959
+ "grad_norm": 0.32859981060028076,
960
+ "learning_rate": 1.999982018053068e-05,
961
+ "loss": 0.11370890140533448,
962
+ "step": 1360
963
+ },
964
+ {
965
+ "epoch": 0.0321108645176412,
966
+ "grad_norm": 0.3100363314151764,
967
+ "learning_rate": 1.999977177553799e-05,
968
+ "loss": 0.1127819538116455,
969
+ "step": 1370
970
+ },
971
+ {
972
+ "epoch": 0.03234525039003274,
973
+ "grad_norm": 0.29671692848205566,
974
+ "learning_rate": 1.9999717608138778e-05,
975
+ "loss": 0.10928714275360107,
976
+ "step": 1380
977
+ },
978
+ {
979
+ "epoch": 0.032579636262424286,
980
+ "grad_norm": 0.3065207302570343,
981
+ "learning_rate": 1.9999657678364252e-05,
982
+ "loss": 0.10931737422943115,
983
+ "step": 1390
984
+ },
985
+ {
986
+ "epoch": 0.03281402213481582,
987
+ "grad_norm": 0.32845526933670044,
988
+ "learning_rate": 1.9999591986248947e-05,
989
+ "loss": 0.1125643253326416,
990
+ "step": 1400
991
+ },
992
+ {
993
+ "epoch": 0.03304840800720737,
994
+ "grad_norm": 0.31998366117477417,
995
+ "learning_rate": 1.999952053183072e-05,
996
+ "loss": 0.11007034778594971,
997
+ "step": 1410
998
+ },
999
+ {
1000
+ "epoch": 0.033282793879598906,
1001
+ "grad_norm": 0.3302232623100281,
1002
+ "learning_rate": 1.9999443315150742e-05,
1003
+ "loss": 0.11075022220611572,
1004
+ "step": 1420
1005
+ },
1006
+ {
1007
+ "epoch": 0.03351717975199045,
1008
+ "grad_norm": 0.3401246964931488,
1009
+ "learning_rate": 1.999936033625352e-05,
1010
+ "loss": 0.10927612781524658,
1011
+ "step": 1430
1012
+ },
1013
+ {
1014
+ "epoch": 0.03375156562438199,
1015
+ "grad_norm": 0.31189054250717163,
1016
+ "learning_rate": 1.999927159518686e-05,
1017
+ "loss": 0.10789297819137574,
1018
+ "step": 1440
1019
+ },
1020
+ {
1021
+ "epoch": 0.03398595149677353,
1022
+ "grad_norm": 0.32660672068595886,
1023
+ "learning_rate": 1.9999177092001907e-05,
1024
+ "loss": 0.10693231821060181,
1025
+ "step": 1450
1026
+ },
1027
+ {
1028
+ "epoch": 0.03422033736916508,
1029
+ "grad_norm": 0.31303200125694275,
1030
+ "learning_rate": 1.9999076826753116e-05,
1031
+ "loss": 0.11288677453994751,
1032
+ "step": 1460
1033
+ },
1034
+ {
1035
+ "epoch": 0.034454723241556615,
1036
+ "grad_norm": 0.3008570075035095,
1037
+ "learning_rate": 1.9998970799498262e-05,
1038
+ "loss": 0.10686779022216797,
1039
+ "step": 1470
1040
+ },
1041
+ {
1042
+ "epoch": 0.03468910911394816,
1043
+ "grad_norm": 0.3439565598964691,
1044
+ "learning_rate": 1.9998859010298447e-05,
1045
+ "loss": 0.1086077332496643,
1046
+ "step": 1480
1047
+ },
1048
+ {
1049
+ "epoch": 0.0349234949863397,
1050
+ "grad_norm": 0.3342174291610718,
1051
+ "learning_rate": 1.9998741459218093e-05,
1052
+ "loss": 0.1085971713066101,
1053
+ "step": 1490
1054
+ },
1055
+ {
1056
+ "epoch": 0.03515788085873124,
1057
+ "grad_norm": 0.3462369740009308,
1058
+ "learning_rate": 1.999861814632493e-05,
1059
+ "loss": 0.11292096376419067,
1060
+ "step": 1500
1061
+ },
1062
+ {
1063
+ "epoch": 0.03539226673112278,
1064
+ "grad_norm": 0.2996586263179779,
1065
+ "learning_rate": 1.9998489071690024e-05,
1066
+ "loss": 0.1069232702255249,
1067
+ "step": 1510
1068
+ },
1069
+ {
1070
+ "epoch": 0.035626652603514324,
1071
+ "grad_norm": 0.30933356285095215,
1072
+ "learning_rate": 1.9998354235387758e-05,
1073
+ "loss": 0.11092567443847656,
1074
+ "step": 1520
1075
+ },
1076
+ {
1077
+ "epoch": 0.03586103847590587,
1078
+ "grad_norm": 0.3398067355155945,
1079
+ "learning_rate": 1.9998213637495824e-05,
1080
+ "loss": 0.10429364442825317,
1081
+ "step": 1530
1082
+ },
1083
+ {
1084
+ "epoch": 0.036095424348297406,
1085
+ "grad_norm": 0.34331822395324707,
1086
+ "learning_rate": 1.999806727809525e-05,
1087
+ "loss": 0.11222647428512574,
1088
+ "step": 1540
1089
+ },
1090
+ {
1091
+ "epoch": 0.03632981022068895,
1092
+ "grad_norm": 0.32439032196998596,
1093
+ "learning_rate": 1.9997915157270367e-05,
1094
+ "loss": 0.10828479528427123,
1095
+ "step": 1550
1096
+ },
1097
+ {
1098
+ "epoch": 0.03656419609308049,
1099
+ "grad_norm": 0.2919040322303772,
1100
+ "learning_rate": 1.9997757275108847e-05,
1101
+ "loss": 0.1067802906036377,
1102
+ "step": 1560
1103
+ },
1104
+ {
1105
+ "epoch": 0.03679858196547203,
1106
+ "grad_norm": 0.3021163046360016,
1107
+ "learning_rate": 1.999759363170166e-05,
1108
+ "loss": 0.1080705165863037,
1109
+ "step": 1570
1110
+ },
1111
+ {
1112
+ "epoch": 0.03703296783786357,
1113
+ "grad_norm": 0.309612512588501,
1114
+ "learning_rate": 1.9997424227143108e-05,
1115
+ "loss": 0.10892333984375,
1116
+ "step": 1580
1117
+ },
1118
+ {
1119
+ "epoch": 0.037267353710255115,
1120
+ "grad_norm": 0.2833057641983032,
1121
+ "learning_rate": 1.9997249061530814e-05,
1122
+ "loss": 0.11133861541748047,
1123
+ "step": 1590
1124
+ },
1125
+ {
1126
+ "epoch": 0.03750173958264666,
1127
+ "grad_norm": 0.32193875312805176,
1128
+ "learning_rate": 1.999706813496572e-05,
1129
+ "loss": 0.11482088565826416,
1130
+ "step": 1600
1131
+ },
1132
+ {
1133
+ "epoch": 0.0377361254550382,
1134
+ "grad_norm": 0.3027825653553009,
1135
+ "learning_rate": 1.9996881447552077e-05,
1136
+ "loss": 0.10923588275909424,
1137
+ "step": 1610
1138
+ },
1139
+ {
1140
+ "epoch": 0.03797051132742974,
1141
+ "grad_norm": 0.321659654378891,
1142
+ "learning_rate": 1.9996688999397473e-05,
1143
+ "loss": 0.10786941051483154,
1144
+ "step": 1620
1145
+ },
1146
+ {
1147
+ "epoch": 0.03820489719982128,
1148
+ "grad_norm": 0.31725117564201355,
1149
+ "learning_rate": 1.99964907906128e-05,
1150
+ "loss": 0.10536646842956543,
1151
+ "step": 1630
1152
+ },
1153
+ {
1154
+ "epoch": 0.038439283072212824,
1155
+ "grad_norm": 0.2907871603965759,
1156
+ "learning_rate": 1.9996286821312282e-05,
1157
+ "loss": 0.10624017715454101,
1158
+ "step": 1640
1159
+ },
1160
+ {
1161
+ "epoch": 0.03867366894460436,
1162
+ "grad_norm": 0.32546359300613403,
1163
+ "learning_rate": 1.9996077091613454e-05,
1164
+ "loss": 0.10742425918579102,
1165
+ "step": 1650
1166
+ },
1167
+ {
1168
+ "epoch": 0.038908054816995906,
1169
+ "grad_norm": 0.2799864411354065,
1170
+ "learning_rate": 1.9995861601637175e-05,
1171
+ "loss": 0.11092619895935059,
1172
+ "step": 1660
1173
+ },
1174
+ {
1175
+ "epoch": 0.03914244068938745,
1176
+ "grad_norm": 0.3120432496070862,
1177
+ "learning_rate": 1.9995640351507623e-05,
1178
+ "loss": 0.10951026678085327,
1179
+ "step": 1670
1180
+ },
1181
+ {
1182
+ "epoch": 0.03937682656177899,
1183
+ "grad_norm": 0.35961198806762695,
1184
+ "learning_rate": 1.9995413341352293e-05,
1185
+ "loss": 0.11031874418258666,
1186
+ "step": 1680
1187
+ },
1188
+ {
1189
+ "epoch": 0.03961121243417053,
1190
+ "grad_norm": 0.3039030134677887,
1191
+ "learning_rate": 1.9995180571301998e-05,
1192
+ "loss": 0.10726022720336914,
1193
+ "step": 1690
1194
+ },
1195
+ {
1196
+ "epoch": 0.03984559830656207,
1197
+ "grad_norm": 0.29744887351989746,
1198
+ "learning_rate": 1.9994942041490878e-05,
1199
+ "loss": 0.11042420864105225,
1200
+ "step": 1700
1201
+ },
1202
+ {
1203
+ "epoch": 0.040079984178953615,
1204
+ "grad_norm": 0.2774674594402313,
1205
+ "learning_rate": 1.9994697752056385e-05,
1206
+ "loss": 0.11003159284591675,
1207
+ "step": 1710
1208
+ },
1209
+ {
1210
+ "epoch": 0.04031437005134515,
1211
+ "grad_norm": 0.3363332748413086,
1212
+ "learning_rate": 1.9994447703139286e-05,
1213
+ "loss": 0.10592604875564575,
1214
+ "step": 1720
1215
+ },
1216
+ {
1217
+ "epoch": 0.0405487559237367,
1218
+ "grad_norm": 0.27715635299682617,
1219
+ "learning_rate": 1.999419189488368e-05,
1220
+ "loss": 0.10486645698547363,
1221
+ "step": 1730
1222
+ },
1223
+ {
1224
+ "epoch": 0.04078314179612824,
1225
+ "grad_norm": 0.3067134916782379,
1226
+ "learning_rate": 1.999393032743697e-05,
1227
+ "loss": 0.10921738147735596,
1228
+ "step": 1740
1229
+ },
1230
+ {
1231
+ "epoch": 0.04101752766851978,
1232
+ "grad_norm": 0.2900826632976532,
1233
+ "learning_rate": 1.9993663000949893e-05,
1234
+ "loss": 0.10426084995269776,
1235
+ "step": 1750
1236
+ },
1237
+ {
1238
+ "epoch": 0.041251913540911324,
1239
+ "grad_norm": 0.2976682186126709,
1240
+ "learning_rate": 1.9993389915576494e-05,
1241
+ "loss": 0.10740503072738647,
1242
+ "step": 1760
1243
+ },
1244
+ {
1245
+ "epoch": 0.04148629941330286,
1246
+ "grad_norm": 0.33628037571907043,
1247
+ "learning_rate": 1.9993111071474138e-05,
1248
+ "loss": 0.10685398578643798,
1249
+ "step": 1770
1250
+ },
1251
+ {
1252
+ "epoch": 0.041720685285694406,
1253
+ "grad_norm": 0.2832029461860657,
1254
+ "learning_rate": 1.999282646880351e-05,
1255
+ "loss": 0.10406830310821533,
1256
+ "step": 1780
1257
+ },
1258
+ {
1259
+ "epoch": 0.041955071158085944,
1260
+ "grad_norm": 0.29356080293655396,
1261
+ "learning_rate": 1.999253610772862e-05,
1262
+ "loss": 0.10723583698272705,
1263
+ "step": 1790
1264
+ },
1265
+ {
1266
+ "epoch": 0.04218945703047749,
1267
+ "grad_norm": 0.29421284794807434,
1268
+ "learning_rate": 1.9992239988416778e-05,
1269
+ "loss": 0.10370919704437256,
1270
+ "step": 1800
1271
+ },
1272
+ {
1273
+ "epoch": 0.04242384290286903,
1274
+ "grad_norm": 0.3112437129020691,
1275
+ "learning_rate": 1.999193811103863e-05,
1276
+ "loss": 0.10596777200698852,
1277
+ "step": 1810
1278
+ },
1279
+ {
1280
+ "epoch": 0.04265822877526057,
1281
+ "grad_norm": 0.29095956683158875,
1282
+ "learning_rate": 1.9991630475768136e-05,
1283
+ "loss": 0.10655193328857422,
1284
+ "step": 1820
1285
+ },
1286
+ {
1287
+ "epoch": 0.042892614647652115,
1288
+ "grad_norm": 0.3111267685890198,
1289
+ "learning_rate": 1.9991317082782572e-05,
1290
+ "loss": 0.1068030595779419,
1291
+ "step": 1830
1292
+ },
1293
+ {
1294
+ "epoch": 0.04312700052004365,
1295
+ "grad_norm": 0.2968071699142456,
1296
+ "learning_rate": 1.9990997932262525e-05,
1297
+ "loss": 0.10419989824295044,
1298
+ "step": 1840
1299
+ },
1300
+ {
1301
+ "epoch": 0.0433613863924352,
1302
+ "grad_norm": 0.2839556336402893,
1303
+ "learning_rate": 1.9990673024391914e-05,
1304
+ "loss": 0.10594829320907592,
1305
+ "step": 1850
1306
+ },
1307
+ {
1308
+ "epoch": 0.043595772264826735,
1309
+ "grad_norm": 0.2657758295536041,
1310
+ "learning_rate": 1.999034235935797e-05,
1311
+ "loss": 0.10265007019042968,
1312
+ "step": 1860
1313
+ },
1314
+ {
1315
+ "epoch": 0.04383015813721828,
1316
+ "grad_norm": 0.2868812084197998,
1317
+ "learning_rate": 1.999000593735123e-05,
1318
+ "loss": 0.10612194538116455,
1319
+ "step": 1870
1320
+ },
1321
+ {
1322
+ "epoch": 0.044064544009609824,
1323
+ "grad_norm": 0.2973288297653198,
1324
+ "learning_rate": 1.9989663758565565e-05,
1325
+ "loss": 0.1020740270614624,
1326
+ "step": 1880
1327
+ },
1328
+ {
1329
+ "epoch": 0.04429892988200136,
1330
+ "grad_norm": 0.33596375584602356,
1331
+ "learning_rate": 1.9989315823198158e-05,
1332
+ "loss": 0.10814613103866577,
1333
+ "step": 1890
1334
+ },
1335
+ {
1336
+ "epoch": 0.044533315754392906,
1337
+ "grad_norm": 0.30708539485931396,
1338
+ "learning_rate": 1.998896213144951e-05,
1339
+ "loss": 0.10821013450622559,
1340
+ "step": 1900
1341
+ },
1342
+ {
1343
+ "epoch": 0.044767701626784444,
1344
+ "grad_norm": 0.306995689868927,
1345
+ "learning_rate": 1.9988602683523427e-05,
1346
+ "loss": 0.10907049179077148,
1347
+ "step": 1910
1348
+ },
1349
+ {
1350
+ "epoch": 0.04500208749917599,
1351
+ "grad_norm": 0.3026825785636902,
1352
+ "learning_rate": 1.998823747962705e-05,
1353
+ "loss": 0.1026546835899353,
1354
+ "step": 1920
1355
+ },
1356
+ {
1357
+ "epoch": 0.045236473371567526,
1358
+ "grad_norm": 0.26199498772621155,
1359
+ "learning_rate": 1.9987866519970832e-05,
1360
+ "loss": 0.10394124984741211,
1361
+ "step": 1930
1362
+ },
1363
+ {
1364
+ "epoch": 0.04547085924395907,
1365
+ "grad_norm": 0.28023838996887207,
1366
+ "learning_rate": 1.9987489804768532e-05,
1367
+ "loss": 0.1040355920791626,
1368
+ "step": 1940
1369
+ },
1370
+ {
1371
+ "epoch": 0.045705245116350615,
1372
+ "grad_norm": 0.3118273913860321,
1373
+ "learning_rate": 1.998710733423724e-05,
1374
+ "loss": 0.10564100742340088,
1375
+ "step": 1950
1376
+ },
1377
+ {
1378
+ "epoch": 0.04593963098874215,
1379
+ "grad_norm": 0.28954198956489563,
1380
+ "learning_rate": 1.9986719108597354e-05,
1381
+ "loss": 0.10645655393600464,
1382
+ "step": 1960
1383
+ },
1384
+ {
1385
+ "epoch": 0.0461740168611337,
1386
+ "grad_norm": 0.33579158782958984,
1387
+ "learning_rate": 1.9986325128072585e-05,
1388
+ "loss": 0.10616481304168701,
1389
+ "step": 1970
1390
+ },
1391
+ {
1392
+ "epoch": 0.046408402733525235,
1393
+ "grad_norm": 0.2787134349346161,
1394
+ "learning_rate": 1.9985925392889974e-05,
1395
+ "loss": 0.10173577070236206,
1396
+ "step": 1980
1397
+ },
1398
+ {
1399
+ "epoch": 0.04664278860591678,
1400
+ "grad_norm": 0.29140785336494446,
1401
+ "learning_rate": 1.9985519903279868e-05,
1402
+ "loss": 0.10528752803802491,
1403
+ "step": 1990
1404
+ },
1405
+ {
1406
+ "epoch": 0.04687717447830832,
1407
+ "grad_norm": 0.2874172329902649,
1408
+ "learning_rate": 1.998510865947593e-05,
1409
+ "loss": 0.10788161754608154,
1410
+ "step": 2000
1411
+ },
1412
+ {
1413
+ "epoch": 0.04711156035069986,
1414
+ "grad_norm": 0.28292685747146606,
1415
+ "learning_rate": 1.998469166171514e-05,
1416
+ "loss": 0.1097710132598877,
1417
+ "step": 2010
1418
+ },
1419
+ {
1420
+ "epoch": 0.047345946223091406,
1421
+ "grad_norm": 0.28411567211151123,
1422
+ "learning_rate": 1.9984268910237797e-05,
1423
+ "loss": 0.10665783882141114,
1424
+ "step": 2020
1425
+ },
1426
+ {
1427
+ "epoch": 0.047580332095482944,
1428
+ "grad_norm": 0.2909773290157318,
1429
+ "learning_rate": 1.9983840405287507e-05,
1430
+ "loss": 0.10285806655883789,
1431
+ "step": 2030
1432
+ },
1433
+ {
1434
+ "epoch": 0.04781471796787449,
1435
+ "grad_norm": 0.29889369010925293,
1436
+ "learning_rate": 1.998340614711121e-05,
1437
+ "loss": 0.10465379953384399,
1438
+ "step": 2040
1439
+ },
1440
+ {
1441
+ "epoch": 0.048049103840266026,
1442
+ "grad_norm": 0.29105401039123535,
1443
+ "learning_rate": 1.998296613595914e-05,
1444
+ "loss": 0.10770035982131958,
1445
+ "step": 2050
1446
+ },
1447
+ {
1448
+ "epoch": 0.04828348971265757,
1449
+ "grad_norm": 0.2655159533023834,
1450
+ "learning_rate": 1.9982520372084855e-05,
1451
+ "loss": 0.10434014797210693,
1452
+ "step": 2060
1453
+ },
1454
+ {
1455
+ "epoch": 0.04851787558504911,
1456
+ "grad_norm": 0.24804545938968658,
1457
+ "learning_rate": 1.9982068855745232e-05,
1458
+ "loss": 0.09603033661842346,
1459
+ "step": 2070
1460
+ },
1461
+ {
1462
+ "epoch": 0.04875226145744065,
1463
+ "grad_norm": 0.27127230167388916,
1464
+ "learning_rate": 1.9981611587200453e-05,
1465
+ "loss": 0.10701797008514405,
1466
+ "step": 2080
1467
+ },
1468
+ {
1469
+ "epoch": 0.0489866473298322,
1470
+ "grad_norm": 0.3288525342941284,
1471
+ "learning_rate": 1.998114856671403e-05,
1472
+ "loss": 0.10488884449005127,
1473
+ "step": 2090
1474
+ },
1475
+ {
1476
+ "epoch": 0.049221033202223735,
1477
+ "grad_norm": 0.2957431375980377,
1478
+ "learning_rate": 1.998067979455277e-05,
1479
+ "loss": 0.10282833576202392,
1480
+ "step": 2100
1481
+ },
1482
+ {
1483
+ "epoch": 0.04945541907461528,
1484
+ "grad_norm": 0.3068355619907379,
1485
+ "learning_rate": 1.998020527098681e-05,
1486
+ "loss": 0.10160216093063354,
1487
+ "step": 2110
1488
+ },
1489
+ {
1490
+ "epoch": 0.04968980494700682,
1491
+ "grad_norm": 0.3347589075565338,
1492
+ "learning_rate": 1.99797249962896e-05,
1493
+ "loss": 0.10400426387786865,
1494
+ "step": 2120
1495
+ },
1496
+ {
1497
+ "epoch": 0.04992419081939836,
1498
+ "grad_norm": 0.30906444787979126,
1499
+ "learning_rate": 1.997923897073789e-05,
1500
+ "loss": 0.10504637956619263,
1501
+ "step": 2130
1502
+ },
1503
+ {
1504
+ "epoch": 0.0501585766917899,
1505
+ "grad_norm": 0.2961674928665161,
1506
+ "learning_rate": 1.9978747194611767e-05,
1507
+ "loss": 0.10290087461471557,
1508
+ "step": 2140
1509
+ },
1510
+ {
1511
+ "epoch": 0.050392962564181444,
1512
+ "grad_norm": 0.27527812123298645,
1513
+ "learning_rate": 1.9978249668194607e-05,
1514
+ "loss": 0.10352675914764405,
1515
+ "step": 2150
1516
+ },
1517
+ {
1518
+ "epoch": 0.05062734843657299,
1519
+ "grad_norm": 0.2558128237724304,
1520
+ "learning_rate": 1.9977746391773117e-05,
1521
+ "loss": 0.10233690738677978,
1522
+ "step": 2160
1523
+ },
1524
+ {
1525
+ "epoch": 0.050861734308964526,
1526
+ "grad_norm": 0.2849492132663727,
1527
+ "learning_rate": 1.9977237365637313e-05,
1528
+ "loss": 0.10237650871276856,
1529
+ "step": 2170
1530
+ },
1531
+ {
1532
+ "epoch": 0.05109612018135607,
1533
+ "grad_norm": 0.2671148478984833,
1534
+ "learning_rate": 1.997672259008052e-05,
1535
+ "loss": 0.10290088653564453,
1536
+ "step": 2180
1537
+ },
1538
+ {
1539
+ "epoch": 0.05133050605374761,
1540
+ "grad_norm": 0.2684539258480072,
1541
+ "learning_rate": 1.997620206539938e-05,
1542
+ "loss": 0.10364431142807007,
1543
+ "step": 2190
1544
+ },
1545
+ {
1546
+ "epoch": 0.05156489192613915,
1547
+ "grad_norm": 0.2842761278152466,
1548
+ "learning_rate": 1.9975675791893848e-05,
1549
+ "loss": 0.10434544086456299,
1550
+ "step": 2200
1551
+ },
1552
+ {
1553
+ "epoch": 0.05179927779853069,
1554
+ "grad_norm": 0.2870532274246216,
1555
+ "learning_rate": 1.997514376986719e-05,
1556
+ "loss": 0.10496325492858886,
1557
+ "step": 2210
1558
+ },
1559
+ {
1560
+ "epoch": 0.052033663670922235,
1561
+ "grad_norm": 0.295663446187973,
1562
+ "learning_rate": 1.997460599962599e-05,
1563
+ "loss": 0.1019209384918213,
1564
+ "step": 2220
1565
+ },
1566
+ {
1567
+ "epoch": 0.05226804954331378,
1568
+ "grad_norm": 0.2890377342700958,
1569
+ "learning_rate": 1.9974062481480136e-05,
1570
+ "loss": 0.1006014347076416,
1571
+ "step": 2230
1572
+ },
1573
+ {
1574
+ "epoch": 0.05250243541570532,
1575
+ "grad_norm": 0.2824719548225403,
1576
+ "learning_rate": 1.9973513215742836e-05,
1577
+ "loss": 0.10610531568527222,
1578
+ "step": 2240
1579
+ },
1580
+ {
1581
+ "epoch": 0.05273682128809686,
1582
+ "grad_norm": 0.2934313714504242,
1583
+ "learning_rate": 1.9972958202730602e-05,
1584
+ "loss": 0.10438885688781738,
1585
+ "step": 2250
1586
+ },
1587
+ {
1588
+ "epoch": 0.0529712071604884,
1589
+ "grad_norm": 0.27515366673469543,
1590
+ "learning_rate": 1.997239744276326e-05,
1591
+ "loss": 0.10107412338256835,
1592
+ "step": 2260
1593
+ },
1594
+ {
1595
+ "epoch": 0.053205593032879944,
1596
+ "grad_norm": 0.28475672006607056,
1597
+ "learning_rate": 1.9971830936163964e-05,
1598
+ "loss": 0.10238564014434814,
1599
+ "step": 2270
1600
+ },
1601
+ {
1602
+ "epoch": 0.05343997890527148,
1603
+ "grad_norm": 0.28128883242607117,
1604
+ "learning_rate": 1.997125868325915e-05,
1605
+ "loss": 0.1000057339668274,
1606
+ "step": 2280
1607
+ },
1608
+ {
1609
+ "epoch": 0.053674364777663026,
1610
+ "grad_norm": 0.2682857811450958,
1611
+ "learning_rate": 1.9970680684378586e-05,
1612
+ "loss": 0.1018843412399292,
1613
+ "step": 2290
1614
+ },
1615
+ {
1616
+ "epoch": 0.05390875065005457,
1617
+ "grad_norm": 0.27407148480415344,
1618
+ "learning_rate": 1.9970096939855347e-05,
1619
+ "loss": 0.10452917814254761,
1620
+ "step": 2300
1621
+ },
1622
+ {
1623
+ "epoch": 0.05414313652244611,
1624
+ "grad_norm": 0.28061622381210327,
1625
+ "learning_rate": 1.996950745002582e-05,
1626
+ "loss": 0.10294064283370971,
1627
+ "step": 2310
1628
+ },
1629
+ {
1630
+ "epoch": 0.05437752239483765,
1631
+ "grad_norm": 0.27093762159347534,
1632
+ "learning_rate": 1.9968912215229697e-05,
1633
+ "loss": 0.10425781011581421,
1634
+ "step": 2320
1635
+ },
1636
+ {
1637
+ "epoch": 0.05461190826722919,
1638
+ "grad_norm": 0.2750425934791565,
1639
+ "learning_rate": 1.9968311235809988e-05,
1640
+ "loss": 0.10463676452636719,
1641
+ "step": 2330
1642
+ },
1643
+ {
1644
+ "epoch": 0.054846294139620735,
1645
+ "grad_norm": 0.2728804349899292,
1646
+ "learning_rate": 1.9967704512113006e-05,
1647
+ "loss": 0.1036148190498352,
1648
+ "step": 2340
1649
+ },
1650
+ {
1651
+ "epoch": 0.05508068001201227,
1652
+ "grad_norm": 0.26616159081459045,
1653
+ "learning_rate": 1.996709204448838e-05,
1654
+ "loss": 0.10782848596572876,
1655
+ "step": 2350
1656
+ },
1657
+ {
1658
+ "epoch": 0.05531506588440382,
1659
+ "grad_norm": 0.28741851449012756,
1660
+ "learning_rate": 1.9966473833289047e-05,
1661
+ "loss": 0.10164358615875244,
1662
+ "step": 2360
1663
+ },
1664
+ {
1665
+ "epoch": 0.05554945175679536,
1666
+ "grad_norm": 0.2606866955757141,
1667
+ "learning_rate": 1.9965849878871254e-05,
1668
+ "loss": 0.1056321382522583,
1669
+ "step": 2370
1670
+ },
1671
+ {
1672
+ "epoch": 0.0557838376291869,
1673
+ "grad_norm": 0.27213457226753235,
1674
+ "learning_rate": 1.9965220181594555e-05,
1675
+ "loss": 0.1027193546295166,
1676
+ "step": 2380
1677
+ },
1678
+ {
1679
+ "epoch": 0.056018223501578444,
1680
+ "grad_norm": 0.2630506753921509,
1681
+ "learning_rate": 1.9964584741821814e-05,
1682
+ "loss": 0.09874136447906494,
1683
+ "step": 2390
1684
+ },
1685
+ {
1686
+ "epoch": 0.05625260937396998,
1687
+ "grad_norm": 0.27048349380493164,
1688
+ "learning_rate": 1.996394355991921e-05,
1689
+ "loss": 0.10071604251861573,
1690
+ "step": 2400
1691
+ },
1692
+ {
1693
+ "epoch": 0.05648699524636153,
1694
+ "grad_norm": 0.2909840941429138,
1695
+ "learning_rate": 1.9963296636256223e-05,
1696
+ "loss": 0.10374481678009033,
1697
+ "step": 2410
1698
+ },
1699
+ {
1700
+ "epoch": 0.056721381118753064,
1701
+ "grad_norm": 0.2548065781593323,
1702
+ "learning_rate": 1.996264397120565e-05,
1703
+ "loss": 0.09739646911621094,
1704
+ "step": 2420
1705
+ },
1706
+ {
1707
+ "epoch": 0.05695576699114461,
1708
+ "grad_norm": 0.25469157099723816,
1709
+ "learning_rate": 1.9961985565143586e-05,
1710
+ "loss": 0.10347042083740235,
1711
+ "step": 2430
1712
+ },
1713
+ {
1714
+ "epoch": 0.05719015286353615,
1715
+ "grad_norm": 0.27947095036506653,
1716
+ "learning_rate": 1.9961321418449443e-05,
1717
+ "loss": 0.09943625926971436,
1718
+ "step": 2440
1719
+ },
1720
+ {
1721
+ "epoch": 0.05742453873592769,
1722
+ "grad_norm": 0.2800036668777466,
1723
+ "learning_rate": 1.996065153150594e-05,
1724
+ "loss": 0.10229953527450561,
1725
+ "step": 2450
1726
+ },
1727
+ {
1728
+ "epoch": 0.057658924608319236,
1729
+ "grad_norm": 0.24105332791805267,
1730
+ "learning_rate": 1.9959975904699096e-05,
1731
+ "loss": 0.10039407014846802,
1732
+ "step": 2460
1733
+ },
1734
+ {
1735
+ "epoch": 0.05789331048071077,
1736
+ "grad_norm": 0.24678143858909607,
1737
+ "learning_rate": 1.995929453841825e-05,
1738
+ "loss": 0.1001626968383789,
1739
+ "step": 2470
1740
+ },
1741
+ {
1742
+ "epoch": 0.05812769635310232,
1743
+ "grad_norm": 0.24058112502098083,
1744
+ "learning_rate": 1.9958607433056035e-05,
1745
+ "loss": 0.1008292555809021,
1746
+ "step": 2480
1747
+ },
1748
+ {
1749
+ "epoch": 0.058362082225493855,
1750
+ "grad_norm": 0.2694501578807831,
1751
+ "learning_rate": 1.9957914589008405e-05,
1752
+ "loss": 0.09967796802520752,
1753
+ "step": 2490
1754
+ },
1755
+ {
1756
+ "epoch": 0.0585964680978854,
1757
+ "grad_norm": 0.27223488688468933,
1758
+ "learning_rate": 1.995721600667461e-05,
1759
+ "loss": 0.10117262601852417,
1760
+ "step": 2500
1761
+ },
1762
+ {
1763
+ "epoch": 0.058830853970276945,
1764
+ "grad_norm": 0.2681121528148651,
1765
+ "learning_rate": 1.9956511686457213e-05,
1766
+ "loss": 0.0976826786994934,
1767
+ "step": 2510
1768
+ },
1769
+ {
1770
+ "epoch": 0.05906523984266848,
1771
+ "grad_norm": 0.28960487246513367,
1772
+ "learning_rate": 1.9955801628762076e-05,
1773
+ "loss": 0.0989774227142334,
1774
+ "step": 2520
1775
+ },
1776
+ {
1777
+ "epoch": 0.05929962571506003,
1778
+ "grad_norm": 0.281716912984848,
1779
+ "learning_rate": 1.995508583399838e-05,
1780
+ "loss": 0.1036555290222168,
1781
+ "step": 2530
1782
+ },
1783
+ {
1784
+ "epoch": 0.059534011587451564,
1785
+ "grad_norm": 0.2516411542892456,
1786
+ "learning_rate": 1.9954364302578598e-05,
1787
+ "loss": 0.09806258678436279,
1788
+ "step": 2540
1789
+ },
1790
+ {
1791
+ "epoch": 0.05976839745984311,
1792
+ "grad_norm": 0.28333738446235657,
1793
+ "learning_rate": 1.9953637034918523e-05,
1794
+ "loss": 0.10205515623092651,
1795
+ "step": 2550
1796
+ },
1797
+ {
1798
+ "epoch": 0.06000278333223465,
1799
+ "grad_norm": 0.26888540387153625,
1800
+ "learning_rate": 1.9952904031437236e-05,
1801
+ "loss": 0.10150589942932128,
1802
+ "step": 2560
1803
+ },
1804
+ {
1805
+ "epoch": 0.06023716920462619,
1806
+ "grad_norm": 0.28422436118125916,
1807
+ "learning_rate": 1.9952165292557143e-05,
1808
+ "loss": 0.09607882499694824,
1809
+ "step": 2570
1810
+ },
1811
+ {
1812
+ "epoch": 0.060471555077017736,
1813
+ "grad_norm": 0.2646063566207886,
1814
+ "learning_rate": 1.9951420818703935e-05,
1815
+ "loss": 0.0985830307006836,
1816
+ "step": 2580
1817
+ },
1818
+ {
1819
+ "epoch": 0.06070594094940927,
1820
+ "grad_norm": 0.2686246931552887,
1821
+ "learning_rate": 1.9950670610306625e-05,
1822
+ "loss": 0.10168079137802125,
1823
+ "step": 2590
1824
+ },
1825
+ {
1826
+ "epoch": 0.06094032682180082,
1827
+ "grad_norm": 0.2655580937862396,
1828
+ "learning_rate": 1.994991466779752e-05,
1829
+ "loss": 0.10096908807754516,
1830
+ "step": 2600
1831
+ },
1832
+ {
1833
+ "epoch": 0.061174712694192356,
1834
+ "grad_norm": 0.3549988269805908,
1835
+ "learning_rate": 1.9949152991612236e-05,
1836
+ "loss": 0.10471076965332031,
1837
+ "step": 2610
1838
+ },
1839
+ {
1840
+ "epoch": 0.0614090985665839,
1841
+ "grad_norm": 0.2887643277645111,
1842
+ "learning_rate": 1.9948385582189695e-05,
1843
+ "loss": 0.09757264256477356,
1844
+ "step": 2620
1845
+ },
1846
+ {
1847
+ "epoch": 0.06164348443897544,
1848
+ "grad_norm": 0.26402509212493896,
1849
+ "learning_rate": 1.994761243997211e-05,
1850
+ "loss": 0.09927420616149903,
1851
+ "step": 2630
1852
+ },
1853
+ {
1854
+ "epoch": 0.06187787031136698,
1855
+ "grad_norm": 0.25216010212898254,
1856
+ "learning_rate": 1.994683356540502e-05,
1857
+ "loss": 0.10259475708007812,
1858
+ "step": 2640
1859
+ },
1860
+ {
1861
+ "epoch": 0.06211225618375853,
1862
+ "grad_norm": 0.2457595318555832,
1863
+ "learning_rate": 1.9946048958937245e-05,
1864
+ "loss": 0.09863827228546143,
1865
+ "step": 2650
1866
+ },
1867
+ {
1868
+ "epoch": 0.062346642056150064,
1869
+ "grad_norm": 0.24635690450668335,
1870
+ "learning_rate": 1.994525862102092e-05,
1871
+ "loss": 0.09740403294563293,
1872
+ "step": 2660
1873
+ },
1874
+ {
1875
+ "epoch": 0.0625810279285416,
1876
+ "grad_norm": 0.2770674228668213,
1877
+ "learning_rate": 1.9944462552111482e-05,
1878
+ "loss": 0.09908512234687805,
1879
+ "step": 2670
1880
+ },
1881
+ {
1882
+ "epoch": 0.06281541380093315,
1883
+ "grad_norm": 0.2662931978702545,
1884
+ "learning_rate": 1.9943660752667667e-05,
1885
+ "loss": 0.10364282131195068,
1886
+ "step": 2680
1887
+ },
1888
+ {
1889
+ "epoch": 0.06304979967332469,
1890
+ "grad_norm": 0.2732428312301636,
1891
+ "learning_rate": 1.9942853223151514e-05,
1892
+ "loss": 0.10122396945953369,
1893
+ "step": 2690
1894
+ },
1895
+ {
1896
+ "epoch": 0.06328418554571623,
1897
+ "grad_norm": 0.2957812547683716,
1898
+ "learning_rate": 1.994203996402837e-05,
1899
+ "loss": 0.09913806319236755,
1900
+ "step": 2700
1901
+ },
1902
+ {
1903
+ "epoch": 0.06351857141810778,
1904
+ "grad_norm": 0.2504137456417084,
1905
+ "learning_rate": 1.9941220975766874e-05,
1906
+ "loss": 0.10005698204040528,
1907
+ "step": 2710
1908
+ },
1909
+ {
1910
+ "epoch": 0.06375295729049932,
1911
+ "grad_norm": 0.2572518587112427,
1912
+ "learning_rate": 1.9940396258838966e-05,
1913
+ "loss": 0.09756886959075928,
1914
+ "step": 2720
1915
+ },
1916
+ {
1917
+ "epoch": 0.06398734316289086,
1918
+ "grad_norm": 0.2348632514476776,
1919
+ "learning_rate": 1.9939565813719906e-05,
1920
+ "loss": 0.09779543280601502,
1921
+ "step": 2730
1922
+ },
1923
+ {
1924
+ "epoch": 0.0642217290352824,
1925
+ "grad_norm": 0.25319570302963257,
1926
+ "learning_rate": 1.993872964088823e-05,
1927
+ "loss": 0.10350990295410156,
1928
+ "step": 2740
1929
+ },
1930
+ {
1931
+ "epoch": 0.06445611490767394,
1932
+ "grad_norm": 0.26977482438087463,
1933
+ "learning_rate": 1.993788774082579e-05,
1934
+ "loss": 0.09852898716926575,
1935
+ "step": 2750
1936
+ },
1937
+ {
1938
+ "epoch": 0.06469050078006548,
1939
+ "grad_norm": 0.27146127820014954,
1940
+ "learning_rate": 1.9937040114017735e-05,
1941
+ "loss": 0.09648056030273437,
1942
+ "step": 2760
1943
+ },
1944
+ {
1945
+ "epoch": 0.06492488665245702,
1946
+ "grad_norm": 0.2358931303024292,
1947
+ "learning_rate": 1.9936186760952508e-05,
1948
+ "loss": 0.09882181882858276,
1949
+ "step": 2770
1950
+ },
1951
+ {
1952
+ "epoch": 0.06515927252484857,
1953
+ "grad_norm": 0.272926390171051,
1954
+ "learning_rate": 1.9935327682121866e-05,
1955
+ "loss": 0.10181998014450074,
1956
+ "step": 2780
1957
+ },
1958
+ {
1959
+ "epoch": 0.06539365839724011,
1960
+ "grad_norm": 0.25403040647506714,
1961
+ "learning_rate": 1.9934462878020844e-05,
1962
+ "loss": 0.09905003309249878,
1963
+ "step": 2790
1964
+ },
1965
+ {
1966
+ "epoch": 0.06562804426963165,
1967
+ "grad_norm": 0.2518831491470337,
1968
+ "learning_rate": 1.99335923491478e-05,
1969
+ "loss": 0.10159070491790771,
1970
+ "step": 2800
1971
+ },
1972
+ {
1973
+ "epoch": 0.06586243014202318,
1974
+ "grad_norm": 0.2731393575668335,
1975
+ "learning_rate": 1.9932716096004375e-05,
1976
+ "loss": 0.0981027603149414,
1977
+ "step": 2810
1978
+ },
1979
+ {
1980
+ "epoch": 0.06609681601441474,
1981
+ "grad_norm": 0.25448721647262573,
1982
+ "learning_rate": 1.9931834119095513e-05,
1983
+ "loss": 0.10141733884811402,
1984
+ "step": 2820
1985
+ },
1986
+ {
1987
+ "epoch": 0.06633120188680627,
1988
+ "grad_norm": 0.2541082203388214,
1989
+ "learning_rate": 1.9930946418929457e-05,
1990
+ "loss": 0.10162540674209594,
1991
+ "step": 2830
1992
+ },
1993
+ {
1994
+ "epoch": 0.06656558775919781,
1995
+ "grad_norm": 0.24208639562129974,
1996
+ "learning_rate": 1.9930052996017748e-05,
1997
+ "loss": 0.10364505052566528,
1998
+ "step": 2840
1999
+ },
2000
+ {
2001
+ "epoch": 0.06679997363158936,
2002
+ "grad_norm": 0.285247802734375,
2003
+ "learning_rate": 1.9929153850875225e-05,
2004
+ "loss": 0.10242388248443604,
2005
+ "step": 2850
2006
+ },
2007
+ {
2008
+ "epoch": 0.0670343595039809,
2009
+ "grad_norm": 0.26962780952453613,
2010
+ "learning_rate": 1.992824898402002e-05,
2011
+ "loss": 0.10025610923767089,
2012
+ "step": 2860
2013
+ },
2014
+ {
2015
+ "epoch": 0.06726874537637244,
2016
+ "grad_norm": 0.24651190638542175,
2017
+ "learning_rate": 1.992733839597357e-05,
2018
+ "loss": 0.09919982552528381,
2019
+ "step": 2870
2020
+ },
2021
+ {
2022
+ "epoch": 0.06750313124876398,
2023
+ "grad_norm": 0.2675532400608063,
2024
+ "learning_rate": 1.9926422087260605e-05,
2025
+ "loss": 0.09946922659873962,
2026
+ "step": 2880
2027
+ },
2028
+ {
2029
+ "epoch": 0.06773751712115553,
2030
+ "grad_norm": 0.2714926302433014,
2031
+ "learning_rate": 1.9925500058409147e-05,
2032
+ "loss": 0.09951109886169433,
2033
+ "step": 2890
2034
+ },
2035
+ {
2036
+ "epoch": 0.06797190299354706,
2037
+ "grad_norm": 0.2618255913257599,
2038
+ "learning_rate": 1.9924572309950523e-05,
2039
+ "loss": 0.0979630172252655,
2040
+ "step": 2900
2041
+ },
2042
+ {
2043
+ "epoch": 0.0682062888659386,
2044
+ "grad_norm": 0.24111329019069672,
2045
+ "learning_rate": 1.992363884241935e-05,
2046
+ "loss": 0.09887909889221191,
2047
+ "step": 2910
2048
+ },
2049
+ {
2050
+ "epoch": 0.06844067473833015,
2051
+ "grad_norm": 0.253558874130249,
2052
+ "learning_rate": 1.9922699656353544e-05,
2053
+ "loss": 0.09874061942100525,
2054
+ "step": 2920
2055
+ },
2056
+ {
2057
+ "epoch": 0.06867506061072169,
2058
+ "grad_norm": 0.2613140046596527,
2059
+ "learning_rate": 1.992175475229431e-05,
2060
+ "loss": 0.102379310131073,
2061
+ "step": 2930
2062
+ },
2063
+ {
2064
+ "epoch": 0.06890944648311323,
2065
+ "grad_norm": 0.24305768311023712,
2066
+ "learning_rate": 1.9920804130786154e-05,
2067
+ "loss": 0.09398667812347412,
2068
+ "step": 2940
2069
+ },
2070
+ {
2071
+ "epoch": 0.06914383235550477,
2072
+ "grad_norm": 0.2555887699127197,
2073
+ "learning_rate": 1.9919847792376883e-05,
2074
+ "loss": 0.10116375684738159,
2075
+ "step": 2950
2076
+ },
2077
+ {
2078
+ "epoch": 0.06937821822789632,
2079
+ "grad_norm": 0.2499028593301773,
2080
+ "learning_rate": 1.991888573761758e-05,
2081
+ "loss": 0.10240293741226196,
2082
+ "step": 2960
2083
+ },
2084
+ {
2085
+ "epoch": 0.06961260410028786,
2086
+ "grad_norm": 0.25849130749702454,
2087
+ "learning_rate": 1.991791796706264e-05,
2088
+ "loss": 0.10004898309707641,
2089
+ "step": 2970
2090
+ },
2091
+ {
2092
+ "epoch": 0.0698469899726794,
2093
+ "grad_norm": 0.2523442804813385,
2094
+ "learning_rate": 1.9916944481269737e-05,
2095
+ "loss": 0.09540088772773743,
2096
+ "step": 2980
2097
+ },
2098
+ {
2099
+ "epoch": 0.07008137584507095,
2100
+ "grad_norm": 0.26065507531166077,
2101
+ "learning_rate": 1.9915965280799853e-05,
2102
+ "loss": 0.10101289749145508,
2103
+ "step": 2990
2104
+ },
2105
+ {
2106
+ "epoch": 0.07031576171746248,
2107
+ "grad_norm": 0.2680308222770691,
2108
+ "learning_rate": 1.9914980366217253e-05,
2109
+ "loss": 0.09616921544075012,
2110
+ "step": 3000
2111
+ }
2112
+ ],
2113
+ "logging_steps": 10,
2114
+ "max_steps": 42665,
2115
+ "num_input_tokens_seen": 0,
2116
+ "num_train_epochs": 1,
2117
+ "save_steps": 500,
2118
+ "stateful_callbacks": {
2119
+ "TrainerControl": {
2120
+ "args": {
2121
+ "should_epoch_stop": false,
2122
+ "should_evaluate": false,
2123
+ "should_log": false,
2124
+ "should_save": true,
2125
+ "should_training_stop": false
2126
+ },
2127
+ "attributes": {}
2128
+ }
2129
+ },
2130
+ "total_flos": 1.3152695723130618e+19,
2131
+ "train_batch_size": 1,
2132
+ "trial_name": null,
2133
+ "trial_params": null
2134
+ }
checkpoint-3000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f080603a35c503394d377f8f66a531d1e0bc0dd8deac36f2f8b83444f6b6c5d3
3
+ size 5777