Add files using upload-large-folder tool

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,95 @@
+---
+base_model: Zyphra/ZAYA1-8B
+license: apache-2.0
+tags:
+  - nvfp4
+  - w4a4
+  - compressed-tensors
+  - quantization
+  - vllm
+  - blackwell
+  - moe
+pipeline_tag: text-generation
+---
+
+# ZAYA1-8B NVFP4 W4A4
+
+**NVFP4 W4A4 inference for Zyphra's ZAYA1-8B on consumer Blackwell (RTX 5070 Ti, SM120)** —
+4-bit weights *and* 4-bit activations running on native CUTLASS FP4 tensor-core kernels,
+at **102.6 tok/s single-stream / 407.4 tok/s batch-8** in under 6 GB of VRAM.
+
+One other public NVFP4 W4A4 ZAYA1 checkpoint exists — built with NVIDIA ModelOpt
+and validated on a 96 GB workstation Blackwell card, with no published accuracy or
+throughput numbers. This checkpoint is, as far as we know, the first built with
+compressed-tensors on a 16 GB **consumer** card, with its own measured accuracy
+and throughput. Full build/debug writeup, quantization scripts, and reproduction
+steps: [github.com/t-timms/zaya1-godspeed](https://github.com/t-timms/zaya1-godspeed).
+
+## Highlights
+
+| Result | Detail |
+|--------|--------|
+| **102.6 tok/s** single / **407.4 tok/s** batch-8 | vLLM + CUDA graphs on RTX 5070 Ti (12.8× over eager mode) |
+| **5.99 GB checkpoint** | 936 Linears in packed NVFP4 W4A4, 384 outlier-sensitive Linears kept BF16 (mixed precision) |
+| **Checkpoint verified healthy** | Budget-forced GPQA-Diamond rises monotonically with reasoning budget — 45.8% → 62.5% at a 12k-token think budget, within CI of Zyphra's BF16 CoT 71.0% |
+| **vLLM SM120 source build** | `TORCH_CUDA_ARCH_LIST=12.0` build enabling `cutlass_scaled_fp4_mm_sm120a` + FP4 group MoE GEMM — kernels that ship in vLLM source but not in wheels |
+
+## Quantization details
+
+| Field | Value |
+|---|---|
+| Base model | `Zyphra/ZAYA1-8B` |
+| Method | compressed-tensors, `float-quantized` |
+| Scheme | NVFP4, `w4a4` (4-bit weights, 4-bit activations) |
+| Group size | 16 (tensor-group strategy) |
+| Ignored / kept BF16 | `lm_head`, router, norm, qkv, cca modules, plus 12 MoE layers with `max_abs > 500` (384 Linears) |
+| Calibration | 977 samples, ARC-mix calibration set, 936 Linears calibrated |
+| Compressor | `NVFP4PackedCompressor` (`weight_packed_dtype=uint8`, `weight_scale_dtype=float8_e4m3fn`) |
+| Built on | RTX 5070 Ti, 16 GB VRAM, CUDA 13.0 |
+
+See `quantization_manifest.json` in this repo for the full machine-readable config,
+including the exact outlier-layer list and mixed-precision threshold.
+
+## Why this is hard
+
+ZAYA1-8B is an 80-layer MoE (760M active / 8.4B total) with Zyphra's CCA
+(compressed convolutional attention) — no stock quantization path works out of
+the box. Notably: W4A4 requires calibrating activation scales (not just
+weights), the compressed-tensors calibration path has a silent NaN-producing
+trap if you calibrate through the fake-quant `nn.Linear.forward` wrapper, and
+the NVFP4 global-scale convention (`2688 / max_abs`, divisor form, block
+scales pre-multiplied) is undocumented — getting it wrong produces silent
+pad-token collapse, not an error. Full root-cause writeup in `RESEARCH.md` on
+the [GitHub repo](https://github.com/t-timms/zaya1-godspeed).
+
+## Usage
+
+This checkpoint requires vLLM built from source with SM120 NVFP4 CUTLASS
+kernels (stock wheels don't include them) — see the
+[reproduce steps](https://github.com/t-timms/zaya1-godspeed#reproduce) in the
+GitHub repo. Inference must run in `bfloat16` (not fp16/fp32).
+
+```bash
+# vLLM built from source with: TORCH_CUDA_ARCH_LIST=12.0
+vllm serve Ttimms/zaya1-8b-nvfp4-w4a4 --dtype bfloat16
+```
+
+## Evaluation
+
+Budget-forced GPQA-Diamond (n=24, paired), using an s1-style harness that caps
+the reasoning trace and scores only the closed answer — stock lm-eval
+harnesses score near-random on this model because ZAYA never closes its
+`<think>` block within a normal budget:
+
+| think budget | GPQA-Diamond | traces self-closing `</think>` |
+|---|---|---|
+| 2,500 | 45.8% | 1/24 |
+| 5,000 | 45.8% | 2/24 |
+| 12,000 | **62.5%** | 9/24 |
+
+The gap to Zyphra's BF16 CoT (71.0%) is the 16 GB context/reasoning-budget
+ceiling on the source hardware — not quantization damage.
+
+## License
+
+Apache 2.0 — matches the ZAYA1-8B upstream license.

chat_template.jinja

@@ -0,0 +1,205 @@
+{% macro render_extra_keys(json_dict, handled_keys) %}
+    {%- if json_dict is mapping %}
+        {%- for json_key in json_dict if json_key not in handled_keys %}
+            {%- if json_dict[json_key] is mapping or (json_dict[json_key] is sequence and json_dict[json_key] is not string) %}
+                {{- '\n<' ~ json_key ~ '>' ~ (json_dict[json_key] | tojson | safe) ~ '</' ~ json_key ~ '>' }}
+            {%- else %}
+                {{-'\n<' ~ json_key ~ '>' ~ (json_dict[json_key] | string) ~ '</' ~ json_key ~ '>' }}
+            {%- endif %}
+        {%- endfor %}
+    {%- endif %}
+{% endmacro %}
+{%- set enable_thinking = enable_thinking if enable_thinking is defined else True %}
+{# TODO: set truncate to true for deployment & agent evals. Keep on for SFT. #}
+{%- set truncate_history_thinking = truncate_history_thinking if truncate_history_thinking is defined else False %}
+{{- bos_token }}
+{%- set ns = namespace(last_user_idx = -1) %}
+{%- set loop_messages = messages %}
+{%- for m in loop_messages %}
+  {%- if m["role"] == "user" %}
+    {%- set ns.last_user_idx = loop.index0 %}
+  {%- endif %}
+{%- endfor %}
+
+{%- if messages[0]["role"] == "system" %}
+    {%- set system_message = messages[0]["content"] %}
+    {%- set loop_messages = messages[1:] %}
+{%- else %}
+    {%- set system_message = "" %}
+    {%- set loop_messages = messages %}
+{%- endif %}
+{%- if not tools is defined %}
+    {%- set tools = [] %}
+{%- endif %}
+{# Recompute last_user_idx relative to loop_messages after handling system #}
+{%- set ns = namespace(last_user_idx = -1) %}
+{%- for m in loop_messages %}
+  {%- if m["role"] == "user" %}
+    {%- set ns.last_user_idx = loop.index0 %}
+  {%- endif %}
+{%- endfor %}
+{%- if system_message is defined %}
+    {{- "<|im_start|>system\n" + system_message }}
+{%- else %}
+    {%- if tools is iterable and tools | length > 0 %}
+        {{- "<|im_start|>system\n" }}
+    {%- endif %}
+{%- endif %}
+{%- if tools is iterable and tools | length > 0 %}
+    {%- if system_message is defined and system_message | length > 0 %}
+        {{- "\n\n" }}
+    {%- endif %}
+    {{- "# Tools\n\nYou have access to the following functions:\n\n" }}
+    {{- "<tools>" }}
+    {%- for tool in tools %}
+        {%- if tool.function is defined %}
+            {%- set tool = tool.function %}
+        {%- endif %}
+        {{- "\n<function>\n<name>" ~ tool.name ~ "</name>" }}
+        {%- if tool.description is defined %}
+            {{- '\n<description>' ~ (tool.description | trim) ~ '</description>' }}
+        {%- endif %}
+        {{- '\n<parameters>' }}
+        {%- if tool.parameters is defined and tool.parameters is mapping and tool.parameters.properties is defined and tool.parameters.properties is mapping %}
+            {%- for param_name, param_fields in tool.parameters.properties|items %}
+                {{- '\n<parameter>' }}
+                {{- '\n<name>' ~ param_name ~ '</name>' }}
+                {%- if param_fields.type is defined %}
+                    {{- '\n<type>' ~ (param_fields.type | string) ~ '</type>' }}
+                {%- endif %}
+                {%- if param_fields.description is defined %}
+                    {{- '\n<description>' ~ (param_fields.description | trim) ~ '</description>' }}
+                {%- endif %}
+                {%- if param_fields.enum is defined %}
+                    {{- '\n<enum>' ~ (param_fields.enum | tojson | safe) ~ '</enum>' }}
+                {%- endif %}
+                {%- set handled_keys = ['name', 'type', 'description', 'enum'] %}
+                {{- render_extra_keys(param_fields, handled_keys) }}
+                {{- '\n</parameter>' }}
+            {%- endfor %}
+        {%- endif %}
+        {% set handled_keys = ['type', 'properties', 'required'] %}
+        {{- render_extra_keys(tool.parameters, handled_keys) }}
+        {%- if tool.parameters is defined and tool.parameters.required is defined %}
+            {{- '\n<required>' ~ (tool.parameters.required | tojson | safe) ~ '</required>' }}
+        {%- endif %}
+        {{- '\n</parameters>' }}
+        {%- set handled_keys = ['type', 'name', 'description', 'parameters'] %}
+        {{- render_extra_keys(tool, handled_keys) }}
+        {{- '\n</function>' }}
+    {%- endfor %}
+    {{- "\n</tools>" }}
+
+    {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<zyphra_tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</zyphra_tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <zyphra_tool_call></zyphra_tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
+{%- endif %}
+
+
+{%- if system_message is defined %}
+    {{- '<|im_end|>\n' }}
+{%- else %}
+    {%- if tools is iterable and tools | length > 0 %}
+        {{- '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+
+{%- for message in loop_messages %}
+    {%- if message.role == "assistant" %}
+        {# Add reasoning content in to content field for unified processing below. #}
+        {%- if message.reasoning_content is defined and message.reasoning_content is string and message.reasoning_content | trim | length > 0 %}
+            {%- set content = "<think>\n" ~ message.reasoning_content ~ "\n</think>\n\n" ~ (message.content | default('', true)) %}
+        {%- else %}
+            {%- set content = message.content | default('', true) %}
+            {%- if content is string -%}
+                {# Allow downstream logic to to take care of broken thought, only handle coherent reasoning here. #}
+                {%- if '<think>' not in content and '</think>' not in content -%}
+                    {%- set content = "<think>\n</think>\n\n" ~ content -%}
+                {%- endif -%}
+            {%- else -%}
+                {%- set content = content -%}
+            {%- endif -%}
+        {%- endif %}
+        {%- if message.tool_calls is defined and message.tool_calls is iterable and message.tool_calls | length > 0 %}
+            {# Assistant message has tool calls. #}
+            {{- '<|im_start|>assistant\n' }}
+                {%- set include_content = not (truncate_history_thinking and loop.index0 < ns.last_user_idx) %}
+                {%- if content is string and content | trim | length > 0 %}
+                    {%- if include_content %}
+                        {{- (content | trim) ~ '\n\n' -}}
+                    {%- else %}
+                        {%- set c = (content | string) %}
+                        {%- if '</think>' in c %}
+                            {# Keep only content after the last closing think. Also generation prompt causes this. #}
+                            {%- set c = c.split('</think>')[-1] %}
+                        {%- elif '<think>' in c %}
+                            {# If <think> was opened but never closed, drop the trailing think segment #}
+                            {%- set c = c.split('<think>')[0] %}
+                        {%- endif %}
+                        {%- set c = "<think>\n</think>\n\n" ~ c | trim %}
+                        {%- if c | length > 0 %}
+                            {{- c ~ '\n' -}}
+                        {%- endif %}
+                    {%- endif %}
+                {%- else %}
+                    {{- "<think>\n</think>\n\n" -}}
+                {%- endif %}
+                {%- for tool_call in message.tool_calls %}
+                    {%- if tool_call.function is defined %}
+                        {%- set tool_call = tool_call.function %}
+                    {%- endif %}
+                    {{- '<zyphra_tool_call>\n<function=' ~ tool_call.name ~ '>\n' -}}
+                        {%- if tool_call.arguments is defined %}
+                            {%- for args_name, args_value in tool_call.arguments|items %}
+                                {{- '<parameter=' ~ args_name ~ '>\n' -}}
+                                    {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
+                                {{- args_value ~ '\n</parameter>\n' -}}
+                            {%- endfor %}
+                        {%- endif %}
+                    {{- '</function>\n</zyphra_tool_call>\n' -}}
+                {%- endfor %}
+                {{- '<|im_end|>\n' }}
+        {%- else %}
+            {# Assistant message doesn't have tool calls. #}
+            {%- if not (truncate_history_thinking and loop.index0 < ns.last_user_idx) %}
+                {{- '<|im_start|>assistant\n' ~ (content | default('', true) | string | trim) ~ '<|im_end|>\n' }}
+            {%- else %}
+                {%- set c = (content | default('', true) | string) %}
+                {%- if '<think>' in c and '</think>' in c %}
+                    {%- set c = "<think>\n</think>\n\n" ~ (c.split('</think>')[-1] | trim) %}
+                {%- endif %}
+                {%- set c = c | trim %}
+                {%- if c | length > 0 %}
+                    {{- '<|im_start|>assistant\n' ~ c ~ '<|im_end|>\n' }}
+                {%- else %}
+                    {{- '<|im_start|>assistant\n<|im_end|>\n' }}
+                {%- endif %}
+            {%- endif %}
+        {%- endif %}
+    {%- elif message.role == "user" or message.role == "system" %}
+        {{- '<|im_start|>' + message.role + '\n' }}
+        {%- set content = message.content | string %}
+        {{- content }}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.previtem and loop.previtem.role != "tool" %}
+            {{- '<|im_start|>user\n' }}
+        {%- endif %}
+        {{- '<zyphra_tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</zyphra_tool_response>\n' }}
+        {%- if not loop.last and loop.nextitem.role != "tool" %}
+            {{- '<|im_end|>\n' }}
+        {%- elif loop.last %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- else %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endfor %}
+
+{%- if add_generation_prompt %}
+    {%- if enable_thinking %}
+        {{- '<|im_start|>assistant\n<think>\n' }}
+    {%- else %}
+        {{- '<|im_start|>assistant\n<think>\n</think>\n\n' }}
+    {%- endif %}
+{%- endif %}

config.json

@@ -0,0 +1,90 @@
+{
+  "activation_func": "swiglu",
+  "activation_func_fp8_input_store": false,
+  "add_bias_linear": false,
+  "architectures": [
+    "ZayaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bias_activation_fusion": true,
+  "bos_token_id": 2,
+  "cca": true,
+  "cca_time0": 2,
+  "cca_time1": 2,
+  "clamp_temp": false,
+  "dtype": "bfloat16",
+  "eos_token_id": 106,
+  "ffn_hidden_size": 4096,
+  "gated_linear_unit": true,
+  "head_dim": 128,
+  "hidden_size": 2048,
+  "kv_channels": 128,
+  "lm_head_bias": false,
+  "mamba_cache_dtype": "float32",
+  "max_position_embeddings": 131072,
+  "model_type": "zaya",
+  "moe_router_topk": 1,
+  "norm_epsilon": 1e-05,
+  "normalization": "RMSNorm",
+  "num_attention_heads": 8,
+  "num_experts": 16,
+  "num_hidden_layers": 80,
+  "num_key_value_heads": 2,
+  "num_query_groups": 2,
+  "pad_token_id": 0,
+  "partial_rotary_factor": 0.5,
+  "residual_in_fp32": true,
+  "rope_parameters": {
+    "partial_rotary_factor": 0.5,
+    "rope_theta": 5000000,
+    "rope_type": "default"
+  },
+  "rope_theta": 5000000,
+  "scale_residual_merge": true,
+  "sliding_window": null,
+  "swa_layers": null,
+  "swa_rotary_base": null,
+  "transformers_version": "4.57.1",
+  "use_cache": true,
+  "vocab_size": 262272,
+  "zaya_high_prec": true,
+  "zaya_mlp_expansion": 256,
+  "zaya_use_eda": true,
+  "zaya_use_mod": true,
+  "quantization_config": {
+    "quant_method": "compressed-tensors",
+    "format": "nvfp4-pack-quantized",
+    "config_groups": {
+      "group_0": {
+        "weights": {
+          "num_bits": 4,
+          "type": "float",
+          "strategy": "tensor_group",
+          "group_size": 16,
+          "symmetric": true,
+          "dynamic": false
+        },
+        "input_activations": {
+          "num_bits": 4,
+          "type": "float",
+          "strategy": "tensor_group",
+          "group_size": 16,
+          "symmetric": true,
+          "dynamic": false
+        },
+        "targets": [
+          "Linear"
+        ]
+      }
+    },
+    "ignore": [
+      "lm_head",
+      "re:.*router.*",
+      "re:.*norm.*",
+      "re:.*qkv.*",
+      "re:.*cca.*",
+      "re:model\\.layers\\.(1|19|31|37|39|65|69|71|73|75|77|79)\\..+"
+    ]
+  }
+}

generation_config.json

@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 2,
+  "eos_token_id": 106,
+  "pad_token_id": 0,
+  "top_k": null,
+  "top_p": null,
+  "transformers_version": "4.57.1"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78702ae84a70df261847fb44e99654c5d3827f80e95d6c3ca422853c159a7487
+size 9457736016

quantization_manifest.json

@@ -0,0 +1,64 @@
+{
+  "model": "Zyphra/ZAYA1-8B",
+  "quantization": {
+    "method": "compressed-tensors",
+    "format": "float-quantized",
+    "scheme": "NVFP4",
+    "scheme_variant": "w4a4",
+    "num_bits_weight": 4,
+    "num_bits_activation": 4,
+    "type": "float",
+    "strategy": "tensor_group",
+    "group_size": 16,
+    "target_modules": [
+      "Linear"
+    ],
+    "ignore": [
+      "lm_head",
+      "re:.*router.*",
+      "re:.*norm.*",
+      "re:.*qkv.*",
+      "re:.*cca.*"
+    ],
+    "weight_packed_dtype": "uint8",
+    "weight_scale_dtype": "float8_e4m3fn",
+    "weight_global_scale_dtype": "float32",
+    "input_global_scale_dtype": "float32",
+    "zero_point": "none (symmetric)",
+    "compressor": "NVFP4PackedCompressor"
+  },
+  "calibration": {
+    "data_path": "data/calibration/arcmix/calibration_data.pt",
+    "num_samples_used": 977,
+    "batch_size": 1,
+    "linears_calibrated": 936
+  },
+  "mixed_precision": {
+    "threshold": 500.0,
+    "outlier_layers": [
+      1,
+      19,
+      31,
+      37,
+      39,
+      65,
+      69,
+      71,
+      73,
+      75,
+      77,
+      79
+    ],
+    "bf16_exempted_modules": 384,
+    "w4a4_compressed_modules": 936,
+    "description": "12 MoE layers where max_abs > 500 kept at BF16 (384 Linears). FusedMoE requires uniform quantization per layer."
+  },
+  "hardware": {
+    "gpu": "NVIDIA GeForce RTX 5070 Ti",
+    "vram": "16 GB",
+    "cuda": "13.0"
+  },
+  "modules_compressed": 936,
+  "output_size_bytes": 9492784787,
+  "dry_run": false
+}

special_tokens_map.json

@@ -0,0 +1,33 @@
+{
+  "boi_token": "<start_of_image>",
+  "bos_token": {
+    "content": "<bos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eoi_token": "<end_of_image>",
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "image_token": "<image_soft_token>",
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:05b019781e631e0bc0bebe417317fe2d2296a0354ab7f206fbb80b446758fd75
+size 33385009