Upload checkpoint-64000

checkpoint-64000/config.json

@@ -0,0 +1,52 @@
+{
+  "architectures": [
+    "Qwen3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 2,
+  "dtype": "float32",
+  "eos_token_id": 3,
+  "head_dim": 64,
+  "hidden_act": "silu",
+  "hidden_size": 512,
+  "initializer_range": 0.02,
+  "intermediate_size": 2048,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 4096,
+  "max_window_layers": 28,
+  "model_type": "qwen3",
+  "num_attention_heads": 8,
+  "num_hidden_layers": 17,
+  "num_key_value_heads": 2,
+  "pad_token_id": 0,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "rope_theta": 20000,
+    "rope_type": "default"
+  },
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.0.0",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 62813
+}

checkpoint-64000/generation_config.json

@@ -0,0 +1,10 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 2,
+  "eos_token_id": 3,
+  "output_attentions": false,
+  "output_hidden_states": false,
+  "pad_token_id": 0,
+  "transformers_version": "5.0.0",
+  "use_cache": false
+}

checkpoint-64000/metadata.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5749e4f27997d8a085f22adc38d023a876f6fed2d11e99e0e898af94c8617ce3
+size 1331

checkpoint-64000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8819cfce59add3a40410c8b4fcd0357d37e9dd4f5617add84dfb16802b8f2c2e
+size 387216312

checkpoint-64000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0d66a7138e94dbab34fbbf6586f26cbcadf6fa14759f04b1a3337602ca5f754
+size 774552395

checkpoint-64000/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6110c1d1bb9a558d96171d9e50c415dd84cc023ad867fc86cb403ae72e193cd
+size 1383

checkpoint-64000/tokenizer_config.json

@@ -0,0 +1,23 @@
+{
+  "backend": "tokenizers",
+  "bos_token": "<|bos|>",
+  "eos_token": "<|eos|>",
+  "extra_special_tokens": [
+    "<|start_of_text|>",
+    "<|end_of_text|>",
+    "<|start_of_audio|>",
+    "<|end_of_audio|>",
+    "<|audio|>",
+    "<|male|>",
+    "<|female|>",
+    "<|other|>",
+    "<|accent_us|>",
+    "<|accent_uk|>",
+    "<|unconditional|>"
+  ],
+  "is_local": true,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<|pad|>",
+  "tokenizer_class": "TokenizersBackend",
+  "unk_token": "<|unk|>"
+}