RL-MPQ Aggressive — 2026-06-11T21:10:07.790343

.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 +1,87 @@
+---
+license: apache-2.0
+base_model: Qwen/Qwen2.5-7B
+pipeline_tag: text-generation
+language:
+  - en
+tags:
+  - qwen2
+  - text-generation
+  - rl-mpq
+  - mixed-precision
+  - quantization
+  - fake-quantization
+  - aggressive
+library_name: transformers
+datasets:
+  - wikitext
+widget:
+  - text: "The capital of France is"
+---
+
+# Qwen 2.5 7B — RL-MPQ Aggressive
+
+Standalone **RL-MPQ** (Reinforcement Learning Mixed-Precision Quantization) checkpoint for the
+**Aggressive** scenario — a quantized variant of
+[Qwen/Qwen2.5-7B](https://huggingface.co/Qwen/Qwen2.5-7B).
+
+| Field | Value |
+|-------|-------|
+| **Base model** | [Qwen/Qwen2.5-7B](https://huggingface.co/Qwen/Qwen2.5-7B) |
+| **Scenario** | Aggressive |
+| **Avg bits / weight** | 3.1429 |
+| **Compression vs FP16** | 5.0909× |
+| **WikiText-2 PPL** | 9.3678 |
+| **Layers** | 28 |
+| **Bit distribution** | `{'3': 24, '4': 4}` |
+| **Format** | Fake-quant FP16 + `rlmpq_policy.json` |
+
+
+## Usage
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+repo = "AvoCahDoe/qwen2-5-7b-rlmpq-aggressive"
+
+model = AutoModelForCausalLM.from_pretrained(repo, torch_dtype="float16")
+tokenizer = AutoTokenizer.from_pretrained(repo)
+```
+
+## Other Qwen 2.5 7B scenarios
+
+| Scenario | Avg bits | Compression | WikiText-2 PPL |
+|----------|----------|-------------|----------------|
+| [Balanced](https://huggingface.co/AvoCahDoe/qwen2-5-7b-rlmpq-balanced) | 3.3929 | 4.7158x | 8.9305 |
+| [Conservative](https://huggingface.co/AvoCahDoe/qwen2-5-7b-rlmpq-conservative) | 3.6786 | 4.3495x | 8.4114 |
+| [Extreme Survival](https://huggingface.co/AvoCahDoe/qwen2-5-7b-rlmpq-extreme-survival) | 2.4643 | 6.4928x | 497.4791 |
+| [High Fidelity](https://huggingface.co/AvoCahDoe/qwen2-5-7b-rlmpq-high-fidelity) | 3.75 | 4.2667x | 8.208 |
+
+Grouped archive (all scenarios in one repo):
+[AvoCahDoe/qwen2-5-7b-rlmpq](https://huggingface.co/AvoCahDoe/qwen2-5-7b-rlmpq)
+
+## Method
+
+1. **Phase 3** — PPO agent assigns per-layer bit widths under the Aggressive reward target.
+2. **Phase 4** — Policy replayed on real weights; WikiText-2 perplexity validates quality.
+3. **Export** — Fake-quantized FP16 weights compatible with Hugging Face Transformers.
+
+## Files
+
+| File | Description |
+|------|-------------|
+| `config.json` | Llama architecture + RL-MPQ metadata |
+| `model.safetensors` | Fake-quantized weights |
+| `rlmpq_policy.json` | Per-layer bit-width policy |
+| `rlmpq_metrics.json` | Validation & PPL summary |
+
+## Citation
+
+```bibtex
+@misc{rlmpq_qwen2_5_7b_aggressive_2026,
+  title  = {RL-MPQ Aggressive: Qwen 2.5 7B Mixed-Precision Quantization},
+  author = {AvoCahDoe},
+  year   = {2026},
+  url    = {https://huggingface.co/AvoCahDoe/qwen2-5-7b-rlmpq-aggressive}
+}
+```

chat_template.jinja

@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\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' }}
+{%- endif %}

config.json

@@ -0,0 +1,79 @@
+{
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "dtype": "bfloat16",
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "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",
+    "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": 131072,
+  "max_window_layers": 28,
+  "model_type": "qwen2",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "pad_token_id": null,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "rope_theta": 1000000.0,
+    "rope_type": "default",
+    "type": "default"
+  },
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "transformers_version": "5.11.0",
+  "use_cache": true,
+  "use_mrope": false,
+  "use_sliding_window": false,
+  "vocab_size": 152064,
+  "base_model": "Qwen/Qwen2.5-7B",
+  "quantization": {
+    "method": "rl-mpq",
+    "format": "fake-quant-fp16",
+    "scenario": "Aggressive",
+    "avg_bits": 3.1429,
+    "description": "Per-layer mixed bit-width via PPO-trained policy (not GPTQ/AWQ)"
+  },
+  "rlmpq": {
+    "framework": "RL-NMP-Model-Quantasation",
+    "scenario": "Aggressive",
+    "scenario_label": "Aggressive",
+    "repo_id": "AvoCahDoe/qwen2-5-7b-rlmpq-aggressive",
+    "grouped_repo": "AvoCahDoe/qwen2-5-7b-rlmpq",
+    "collection_title": "RL-MPQ \u2014 Qwen 2.5 7B"
+  }
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": false,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048,
+  "transformers_version": "5.11.0"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dde7f689c1f870dfb8b32cc4558858324c70e36b27429dfc2846b829cd1e3b4f
+size 15231272152

rlmpq_metrics.json

@@ -0,0 +1,26 @@
+{
+  "framework": "RL-NMP-Model-Quantasation",
+  "method": "RL-MPQ",
+  "repo_id": "AvoCahDoe/qwen2-5-7b-rlmpq-aggressive",
+  "base_model": "Qwen/Qwen2.5-7B",
+  "model_slug": "qwen2_5_7b",
+  "scenario": "Aggressive",
+  "scenario_label": "Aggressive",
+  "avg_bits_per_weight": 3.1429,
+  "compression_vs_fp16": 5.0909,
+  "wikitext2_perplexity": 9.3678,
+  "bit_distribution": {
+    "3": 24,
+    "4": 4
+  },
+  "quantization": {
+    "type": "fake-quant-fp16",
+    "group_size": 128,
+    "scheme": "per-layer asymmetric group-wise",
+    "packed_format": false
+  },
+  "exported_at": "2026-06-11T21:10:06.436057",
+  "policy_source": "/workspace/RL-NMP-Model-Quantasation/phase3/models/qwen2_5_7b/results/Aggressive_policy.json",
+  "policy_path": "/workspace/RL-NMP-Model-Quantasation/phase3/models/qwen2_5_7b/results/Aggressive_policy.json",
+  "training_run": "20260610_202902"
+}

rlmpq_policy.json

@@ -0,0 +1,58 @@
+{
+  "slug": "qwen2_5_7b",
+  "hf_id": "Qwen/Qwen2.5-7B",
+  "scenario": "Aggressive",
+  "params": {
+    "target_bits": 2.75,
+    "lambda_lag": 30,
+    "lambda_mse": 1.2,
+    "lambda_bit": 0.85
+  },
+  "policy": [
+    4,
+    4,
+    4,
+    3,
+    4,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3
+  ],
+  "num_layers": 28,
+  "avg_bits": 3.1429,
+  "total_reward": -28.020284,
+  "compression_ratio": 5.0909,
+  "approx_size_gb": 2.75,
+  "bit_distribution": {
+    "3": 24,
+    "4": 4
+  },
+  "avg_mse_per_bit": {
+    "4": 1.165e-05,
+    "3": 5.632e-05
+  },
+  "validation_elapsed_s": 0.0479,
+  "policy_source": "/workspace/RL-NMP-Model-Quantasation/phase3/models/qwen2_5_7b/results/Aggressive_policy.json",
+  "policy_path": "/workspace/RL-NMP-Model-Quantasation/phase3/models/qwen2_5_7b/results/Aggressive_policy.json",
+  "training_run": "20260610_202902"
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8
+size 11421892

tokenizer_config.json

@@ -0,0 +1,30 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "local_files_only": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}