Upload MiniCPM5-1B Quanto FP8 quantized weights and model card

README.md

@@ -0,0 +1,105 @@
+---
+language:
+- ja
+- en
+- zh
+license: apache-2.0
+library_name: optimum-quanto
+base_model: openbmb/MiniCPM5-1B
+tags:
+- quantized
+- fp8
+- float8
+- e4m3fn
+- optimum-quanto
+- minicpm
+- rtx-4070-ti
+- edge-ai
+- text-generation
+- conversational
+pipeline_tag: text-generation
+---
+
+# MiniCPM5-1B-Quanto-FP8
+
+This repository provides **`openbmb/MiniCPM5-1B` quantized to FP8 (Float8 `e4m3fn`)** using `optimum-quanto`.
+
+It is optimized for 4th Generation Tensor Cores (NVIDIA Ada Lovelace RTX 40xx series, Hopper H100, Blackwell B200) to deliver **ultra-low time-to-first-token (TTFT 76ms)**, **1.78x higher decode throughput (28.3 tok/s)**, and a **compact 1.4GB VRAM footprint** without any loss in reasoning, algebra, or anti-hallucination accuracy compared to native `bfloat16`.
+
+- **Base Model**: [`openbmb/MiniCPM5-1B`](https://huggingface.co/openbmb/MiniCPM5-1B) (1.16B parameters, 128k context, LlamaForCausalLM)
+- **Quantization Method**: `optimum-quanto` FP8 (`weights=qfloat8_e4m3fn`, `activations=None`)
+- **Model Weight Size**: 1.22 GB (`model.safetensors`)
+- **Benchmark Evaluation Dataset**: [`aoiandroid/minicpm5-1b-quantization-benchmark`](https://huggingface.co/datasets/aoiandroid/minicpm5-1b-quantization-benchmark)
+
+---
+
+## Performance Benchmarks (NVIDIA RTX 4070 Ti 12GB)
+
+Measured across 8 authentic conversational, logical, mathematical, and structural scenarios:
+
+| Metric | bfloat16 (Native) | BitsAndBytes 4bit (NF4) | **Quanto FP8 (This Model)** | Quanto INT4 |
+| :--- | :--- | :--- | :--- | :--- |
+| **Model Weight VRAM** | 2071.1 MB (2.02 GB) | 1108.3 MB (1.08 GB) | **1425.0 MB (1.39 GB)** | 1128.8 MB (1.10 GB) |
+| **Peak VRAM During Inference** | 2109.4 MB | 1148.9 MB | **1478.7 MB** | 1166.3 MB |
+| **Time-To-First-Token (TTFT)** | 135.2 ms | 193.3 ms | **76.4 ms (43.6% faster)** | 67.0 ms |
+| **Decode Speed (TPS)** | 15.9 tok/s | 11.5 tok/s | **28.3 tok/s (1.78x faster)** | 32.5 tok/s |
+| **Overall Speed (TPS)** | 15.8 tok/s | 11.4 tok/s | **28.1 tok/s** | 32.3 tok/s |
+| **Reasoning Accuracy (Math/Logic)** | 100% (Full accuracy) | Degraded | **100% (Zero loss vs BF16)** | Severely degraded |
+| **Anti-Hallucination Rejection** | Perfect | Unstable | **Perfect** | Infinite loop |
+
+### Why FP8 Outperforms BitsAndBytes 4bit:
+1. **Hardware Tensor Core Execution**: On Ada Lovelace GPUs, FP8 GEMM is executed directly on 4th-gen Tensor Cores.
+2. **Elimination of Dequantization Overhead**: Traditional 4-bit (BitsAndBytes NF4) suffers from dynamic dequantization kernel overhead before every GEMM operation, causing an inverted slowdown on small (1B) models.
+3. **Zero Accuracy Loss**: Float8 `e4m3fn` preserves dynamic range across attention projections, preventing the logic collapse observed in INT4.
+
+---
+
+## Quickstart Inference Code
+
+### Installation
+```bash
+pip install torch transformers accelerate optimum-quanto
+```
+
+### Direct Loading with `optimum.quanto`
+```python
+import torch
+from optimum.quanto import QuantizedModelForCausalLM
+from transformers import AutoTokenizer
+
+model_id = "aoiandroid/MiniCPM5-1B-Quanto-FP8"
+
+# 1. Load model and move to CUDA
+model = QuantizedModelForCausalLM.from_pretrained(model_id, trust_remote_code=True)
+model.to("cuda")
+
+# 2. Load tokenizer
+tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
+
+# 3. Format prompt using chat template
+prompt = "こんにちは！自己紹介をしてください。"
+messages = [{"role": "user", "content": prompt}]
+input_text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+inputs = tokenizer(input_text, return_tensors="pt").to("cuda")
+inputs.pop("token_type_ids", None)
+
+# 4. Generate with high-speed FP8 Tensor Cores
+with torch.no_grad():
+    output_tokens = model.generate(
+        **inputs,
+        max_new_tokens=256,
+        do_sample=True,
+        temperature=0.7,
+        top_p=0.8
+    )
+
+response = tokenizer.decode(output_tokens[0][inputs.input_ids.shape[1]:], skip_special_tokens=True)
+print(response)
+```
+
+---
+
+## Evaluation Data and Reproducibility
+
+For full raw benchmark output logs (JSON) and evaluation scripts across all 8 test cases, visit the evaluation dataset repository:
+- [`aoiandroid/minicpm5-1b-quantization-benchmark`](https://huggingface.co/datasets/aoiandroid/minicpm5-1b-quantization-benchmark)

chat_template.jinja

@@ -0,0 +1,179 @@
+{{- bos_token }}{%- if tools %}
+    {%- set tool_definitions %}
+        {{- "# Tools\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+        {%- for tool in tools %}
+            {{- "\n" }}
+            {{- tool | tojson(ensure_ascii=False) }}
+        {%- endfor %}
+        {{- '\n</tools>\n\nTool usage guidelines:\n- You may call zero or more functions. If no function calls are needed, just answer normally and do not include any <function ... </function>.\n- When calling a function, return an XML object within <function ... </function> using:\n<function name="function-name"><param name="param-name">param-value</param></function>\n- param-value may be multi-line. If it contains <, & or newline characters, wrap it in a CDATA block: <param name="param-name"><![CDATA[...multi-line value...]]></param>' }}
+    {%- endset %}
+    
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {%- if '<tool_def_sep>' in messages[0].content %}
+            {{- messages[0].content.replace('<tool_def_sep>', tool_definitions) }}
+        {%- else %}
+            {{- messages[0].content + '\n\n' + tool_definitions }}
+        {%- endif %}
+    {%- else %}
+        {{- tool_definitions.lstrip() }}
+    {%- endif %}
+    {{- '<|im_end|>\n' }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if message.content is string %}
+        {%- set content = message.content %}
+    {%- else %}
+        {%- set content = '' %}
+    {%- endif %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is string %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in content %}
+                {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+                {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        
+        {%- if message.tool_calls %}
+            {%- set content_parts = content.split('<tool_sep>') %}
+            {%- set processed_content = content_parts[0] %}
+            {%- set tool_calls_count = message.tool_calls|length %}
+            {%- set tool_sep_count = content_parts|length - 1 %}
+            {%- set min_count = [tool_calls_count, tool_sep_count]|min %}
+            
+            {%- for i in range(1, content_parts|length) %}
+                {%- set tool_index = i - 1 %}
+                {%- if tool_index < tool_calls_count %}
+                    {%- set tool_call = message.tool_calls[tool_index] %}
+                    {%- if tool_call.function %}
+                        {%- set tool_call = tool_call.function %}
+                    {%- endif %}
+                    {%- set single_tool_xml %}
+                        {{- '<function name="' ~ tool_call.name ~ '">' }}
+                        {%- if tool_call.arguments %}
+                            {%- set args_dict = tool_call.arguments %}
+                            {%- for param_name, param_value in args_dict.items() %}
+                                {{- '<param name="' ~ param_name ~ '">' }}
+                                {%- if param_value is string and ('<' in param_value or '&' in param_value or '\n' in param_value) %}
+                                    {{- '<![CDATA[' + param_value + ']]>' }}
+                                {%- else %}
+                                    {{- param_value }}
+                                {%- endif %}
+                                {{- '</param>' }}
+                            {%- endfor %}
+                        {%- endif %}
+                        {{- '</function>' }}
+                    {%- endset %}
+                    {%- set processed_content = processed_content + single_tool_xml + content_parts[i] %}
+                {%- else %}
+                    {%- set processed_content = processed_content + content_parts[i] %}
+                {%- endif %}
+            {%- endfor %}
+            
+            {%- if tool_calls_count > tool_sep_count %}
+                {%- for remaining_index in range(tool_sep_count, tool_calls_count) %}
+                    {%- set tool_call = message.tool_calls[remaining_index] %}
+                    {%- if tool_call.function %}
+                        {%- set tool_call = tool_call.function %}
+                    {%- endif %}
+                    {%- set remaining_tool_xml %}
+                        {{- '<function name="' ~ tool_call.name ~ '">' }}
+                        {%- if tool_call.arguments %}
+                            {%- set args_dict = tool_call.arguments %}
+                            {%- for param_name, param_value in args_dict.items() %}
+                                {{- '<param name="' ~ param_name ~ '">' }}
+                                {%- if param_value is string and ('<' in param_value or '&' in param_value or '\n' in param_value) %}
+                                    {{- '<![CDATA[' + param_value + ']]>' }}
+                                {%- else %}
+                                    {{- param_value }}
+                                {%- endif %}
+                                {{- '</param>' }}
+                            {%- endfor %}
+                        {%- endif %}
+                        {{- '</function>' }}
+                    {%- endset %}
+                    {%- set processed_content = processed_content + remaining_tool_xml %}
+                {%- endfor %}
+            {%- endif %}
+            
+            {%- set content = processed_content %}
+        {%- endif %}
+        
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if reasoning_content %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        
+        {%- if message.tool_calls and not has_tool_sep %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<function name="' ~ tool_call.name ~ '">' }}
+                {%- if tool_call.arguments %}
+                    {%- set args_dict = tool_call.arguments %}
+                    {%- for param_name, param_value in args_dict.items() %}
+                        {{- '<param name="' ~ param_name ~ '">' }}
+                        {%- if param_value is string and ('<' in param_value or '&' in param_value or '\n' in param_value) %}
+                            {{- '<![CDATA[' + param_value + ']]>' }}
+                        {%- else %}
+                            {{- param_value }}
+                        {%- endif %}
+                        {{- '</param>' }}
+                    {%- endfor %}
+                {%- endif %}
+                {{- '</function>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {%- if message.content is string %}
+            {{- content }}
+        {%- else %}
+            {{- message.content | tojson(ensure_ascii=False) }}
+        {%- endif %}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined %}
+        {%- if enable_thinking is false %}
+            {{- '<think>\n\n</think>\n\n' }}
+        {%- elif enable_thinking is true %}
+            {{- '<think>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endif %}

config.json

@@ -0,0 +1,33 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 0,
+  "dtype": "bfloat16",
+  "eos_token_id": [
+    1,
+    130073
+  ],
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "initializer_range": 0.02,
+  "intermediate_size": 4608,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 2,
+  "pad_token_id": 1,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 5000000,
+  "tie_word_embeddings": false,
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "vocab_size": 130560
+}

generation_config.json

@@ -0,0 +1,13 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 0,
+  "do_sample": true,
+  "eos_token_id": [
+    1,
+    130073
+  ],
+  "pad_token_id": 1,
+  "temperature": 0.9,
+  "top_p": 0.95,
+  "transformers_version": "4.57.6"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b705049e5b1ad29f292a9427f8209d8a10f921ad629923a35a7375fca0a54854
+size 1282306460

quanto_qmap.json

@@ -0,0 +1,678 @@
+{
+    "model.layers.0.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.0.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.0.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.0.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.0.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.0.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.0.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.1.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.1.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.1.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.1.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.1.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.1.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.1.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.2.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.2.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.2.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.2.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.2.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.2.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.2.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.3.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.3.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.3.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.3.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.3.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.3.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.3.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.4.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.4.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.4.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.4.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.4.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.4.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.4.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.5.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.5.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.5.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.5.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.5.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.5.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.5.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.6.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.6.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.6.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.6.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.6.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.6.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.6.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.7.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.7.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.7.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.7.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.7.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.7.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.7.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.8.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.8.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.8.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.8.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.8.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.8.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.8.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.9.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.9.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.9.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.9.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.9.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.9.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.9.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.10.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.10.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.10.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.10.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.10.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.10.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.10.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.11.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.11.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.11.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.11.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.11.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.11.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.11.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.12.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.12.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.12.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.12.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.12.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.12.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.12.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.13.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.13.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.13.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.13.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.13.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.13.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.13.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.14.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.14.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.14.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.14.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.14.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.14.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.14.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.15.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.15.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.15.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.15.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.15.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.15.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.15.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.16.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.16.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.16.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.16.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.16.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.16.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.16.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.17.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.17.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.17.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.17.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.17.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.17.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.17.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.18.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.18.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.18.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.18.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.18.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.18.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.18.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.19.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.19.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.19.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.19.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.19.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.19.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.19.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.20.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.20.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.20.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.20.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.20.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.20.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.20.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.21.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.21.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.21.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.21.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.21.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.21.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.21.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.22.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.22.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.22.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.22.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.22.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.22.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.22.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.23.self_attn.q_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.23.self_attn.k_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.23.self_attn.v_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.23.self_attn.o_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.23.mlp.gate_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.23.mlp.up_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "model.layers.23.mlp.down_proj": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    },
+    "lm_head": {
+        "weights": "qfloat8_e4m3fn",
+        "activations": "none"
+    }
+}

special_tokens_map.json

@@ -0,0 +1,30 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}