add model

.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 +1,207 @@
+---
+tags:
+- w8a8
+- int8
+- vllm
+license: apache-2.0
+license_link: https://huggingface.co/Qwen/QwQ-32B-Preview/blob/main/LICENSE
+language:
+  - en
+base_model: Qwen/Qwen2.5-32B-Instruct
+tags:
+  - chat
+library_name: transformers
+---
+
+# QwQ-32B-Preview-quantized.w8a8
+
+## Model Overview
+- **Model Architecture:** QwQ-32B-Preview
+  - **Input:** Text
+  - **Output:** Text
+- **Model Optimizations:**
+  - **Weight quantization:** INT8
+  - **Activation quantization:** INT8
+- **Release Date:** 3/1/2025
+- **Version:** 1.0
+- **Model Developers:** Neural Magic
+
+Quantized version of [QwQ-32B-Preview](https://huggingface.co/Qwen/QwQ-32B-Preview).
+It achieves an average score of 76.49 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 77.20.
+
+### Model Optimizations
+
+This model was obtained by quantizing the weights and activations of [QwQ-32B-Preview](https://huggingface.co/Qwen/QwQ-32B-Preview) to INT8 data type, ready for inference with vLLM >= 0.5.2.
+This optimization reduces the number of bits per parameter from 16 to 8, reducing the disk size and GPU memory requirements by approximately 50%. Only the weights and activations of the linear operators within transformers blocks are quantized. 
+
+## Deployment
+
+### Use with vLLM
+
+This model can be deployed efficiently using the [vLLM](https://docs.vllm.ai/en/latest/) backend, as shown in the example below.
+
+```python
+from transformers import AutoTokenizer
+from vllm import LLM, SamplingParams
+
+max_model_len, tp_size = 4096, 1
+model_name = "neuralmagic-ent/QwQ-32B-Preview-quantized.w8a8"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+llm = LLM(model=model_name, tensor_parallel_size=tp_size, max_model_len=max_model_len, trust_remote_code=True)
+sampling_params = SamplingParams(temperature=0.3, max_tokens=256, stop_token_ids=[tokenizer.eos_token_id])
+
+messages_list = [
+    [{"role": "user", "content": "Who are you? Please respond in pirate speak!"}],
+]
+
+prompt_token_ids = [tokenizer.apply_chat_template(messages, add_generation_prompt=True) for messages in messages_list]
+
+outputs = llm.generate(prompt_token_ids=prompt_token_ids, sampling_params=sampling_params)
+
+generated_text = [output.outputs[0].text for output in outputs]
+print(generated_text)
+```
+
+vLLM also supports OpenAI-compatible serving. See the [documentation](https://docs.vllm.ai/en/latest/) for more details.
+
+## Creation
+
+This model was created with [llm-compressor](https://github.com/vllm-project/llm-compressor) by running the code snippet below with the following arguments:
+
+
+```bash
+python quantize.py --model_path Qwen/QwQ-32B-Preview --quant_path "output_dir/QwQ-32B-Preview-quantized.w8a8" --calib_size 1024 --dampening_frac 0.1 --observer mse 
+```
+
+
+```python
+from datasets import load_dataset
+from transformers import AutoTokenizer
+from llmcompressor.modifiers.quantization import GPTQModifier
+from llmcompressor.transformers import SparseAutoModelForCausalLM, oneshot, apply
+import argparse
+from compressed_tensors.quantization import QuantizationScheme, QuantizationArgs, QuantizationType, QuantizationStrategy
+
+
+parser = argparse.ArgumentParser()
+parser.add_argument('--model_path', type=str)
+parser.add_argument('--quant_path', type=str)
+parser.add_argument('--calib_size', type=int, default=256)
+parser.add_argument('--dampening_frac', type=float, default=0.1) 
+parser.add_argument('--observer', type=str, default="minmax")
+args = parser.parse_args()
+
+model = SparseAutoModelForCausalLM.from_pretrained(
+    args.model_path,
+    device_map="auto",
+    torch_dtype="auto",
+    use_cache=False,
+    trust_remote_code=True,
+)
+tokenizer = AutoTokenizer.from_pretrained(args.model_path)
+
+NUM_CALIBRATION_SAMPLES = args.calib_size
+DATASET_ID = "garage-bAInd/Open-Platypus"
+DATASET_SPLIT = "train"
+ds = load_dataset(DATASET_ID, split=DATASET_SPLIT)
+ds = ds.shuffle(seed=42).select(range(NUM_CALIBRATION_SAMPLES))
+
+def preprocess(example):
+    concat_txt = example["instruction"] + "\n" + example["output"]
+    return {"text": concat_txt}
+
+ds = ds.map(preprocess)
+
+def tokenize(sample):
+    return tokenizer(
+        sample["text"],
+        padding=False,
+        truncation=False,
+        add_special_tokens=True,
+    )
+
+
+ds = ds.map(tokenize, remove_columns=ds.column_names)
+
+recipe = [
+    GPTQModifier(
+        targets=["Linear"],
+        ignore=["lm_head"],
+        scheme="W8A8",
+        dampening_frac=args.dampening_frac,
+        observer=args.observer,
+    )
+]
+oneshot(
+    model=model,
+    dataset=ds,
+    recipe=recipe,
+    num_calibration_samples=args.calib_size,
+    max_seq_length=8192,
+)
+
+# Save to disk compressed.
+SAVE_DIR = args.quant_path
+model.save_pretrained(SAVE_DIR, save_compressed=True)
+tokenizer.save_pretrained(SAVE_DIR)
+```
+
+## Evaluation
+
+The model was evaluated on OpenLLM Leaderboard [V1](https://huggingface.co/spaces/open-llm-leaderboard-old/open_llm_leaderboard) and [V2](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard#/), using the following commands:
+
+OpenLLM Leaderboard V1:
+```
+lm_eval \
+  --model vllm \
+  --model_args pretrained="neuralmagic-ent/QwQ-32B-Preview-quantized.w8a8",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=1,gpu_memory_utilization=0.8,enable_chunked_prefill=True,trust_remote_code=True \
+  --tasks openllm \
+  --write_out \
+  --batch_size auto \
+  --output_path output_dir \
+  --show_config
+```
+
+OpenLLM Leaderboard V2:
+```
+lm_eval \
+  --model vllm \
+  --model_args pretrained="neuralmagic-ent/QwQ-32B-Preview-quantized.w8a8",dtype=auto,add_bos_token=False,max_model_len=4096,tensor_parallel_size=1,gpu_memory_utilization=0.8,enable_chunked_prefill=True,trust_remote_code=True \
+  --apply_chat_template \
+  --fewshot_as_multiturn \
+  --tasks leaderboard \
+  --write_out \
+  --batch_size auto \
+  --output_path output_dir \
+  --show_config
+
+```
+
+### Accuracy
+
+#### OpenLLM Leaderboard V1 evaluation scores
+
+| Metric                                   | Qwen/QwQ-32B-Preview             | neuralmagic-ent/QwQ-32B-Preview-quantized.w8a8 |
+|-----------------------------------------|:---------------------------------:|:-------------------------------------------:|
+| ARC-Challenge (Acc-Norm, 25-shot)       | 70.73                            | 70.73                                       |
+| GSM8K (Strict-Match, 5-shot)            | 83.09                            | 79.91                                       |
+| HellaSwag (Acc-Norm, 10-shot)           | 85.77                            | 85.75                                       |
+| MMLU (Acc, 5-shot)                      | 82.67                            | 82.24                                       |
+| TruthfulQA (MC2, 0-shot)                | 60.88                            | 59.18                                       |
+| Winogrande (Acc, 5-shot)                | 80.03                            | 81.14                                       |
+| **Average Score**                       | **77.20**                        | **76.49**                                   |
+| **Recovery**                            | **100.00**                       | **99.08**                                   |
+
+#### OpenLLM Leaderboard V2 evaluation scores
+
+| Metric                                                   | Qwen/QwQ-32B-Preview             | neuralmagic-ent/QwQ-32B-Preview-quantized.w8a8 |
+|---------------------------------------------------------|:---------------------------------:|:-------------------------------------------:|
+| IFEval (Inst-and-Prompt Level Strict Acc, 0-shot)       | 42.34                            | 43.49                                       |
+| BBH (Acc-Norm, 3-shot)                                  | 53.03                            | 52.95                                       |
+| Math-Hard (Exact-Match, 4-shot)                         | 21.15                            | 22.36                                       |
+| GPQA (Acc-Norm, 0-shot)                                 | 2.97                             | 3.5                                        |
+| MUSR (Acc-Norm, 0-shot)                                 | 9.57                             | 10.87                                       |
+| MMLU-Pro (Acc, 5-shot)                                  | 52.00                            | 51.4                                       |
+| **Average Score**                                       | **30.18**                        | **30.76**                                   |
+| **Recovery**                                            | **100.00**                       | **101.92**                                   |
+

added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

config.json

@@ -0,0 +1,71 @@
+{
+  "_name_or_path": "Qwen/QwQ-32B-Preview",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 5120,
+  "initializer_range": 0.02,
+  "intermediate_size": 27648,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 64,
+  "model_type": "qwen2",
+  "num_attention_heads": 40,
+  "num_hidden_layers": 64,
+  "num_key_value_heads": 8,
+  "quantization_config": {
+    "config_groups": {
+      "group_0": {
+        "input_activations": {
+          "actorder": null,
+          "block_structure": null,
+          "dynamic": true,
+          "group_size": null,
+          "num_bits": 8,
+          "observer": null,
+          "observer_kwargs": {},
+          "strategy": "token",
+          "symmetric": true,
+          "type": "int"
+        },
+        "output_activations": null,
+        "targets": [
+          "Linear"
+        ],
+        "weights": {
+          "actorder": null,
+          "block_structure": null,
+          "dynamic": false,
+          "group_size": null,
+          "num_bits": 8,
+          "observer": "minmax",
+          "observer_kwargs": {},
+          "strategy": "channel",
+          "symmetric": true,
+          "type": "int"
+        }
+      }
+    },
+    "format": "int-quantized",
+    "global_compression_ratio": 1.462761855914978,
+    "ignore": [
+      "lm_head"
+    ],
+    "kv_cache_scheme": null,
+    "quant_method": "compressed-tensors",
+    "quantization_status": "compressed"
+  },
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.47.1",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

generation_config.json

@@ -0,0 +1,13 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.47.1"
+}

model-00001-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de0111c56f51428f9beda9d632521bb9ce79a6fac84a67558d32761cb2a412d3
+size 4997762496

model-00002-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cbc72d4ddd26edf5e1a65b28a21589312e881fb6fc4b1bbae7b149965b89cad2
+size 4914421312

model-00003-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ef88f31d10359ebf3d940b873d271723d5a0d41fcfb23edbbdc4b14d8e978d6
+size 4877701896

model-00004-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ac4a9378b7116b360ccb67385be46dc0e896c717c0b3bc11a86dad94eaf9c02
+size 4877701896

model-00005-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:085323cfeb249569a23ee657411a7d3f48464273decc8d50ae08391bba24329c
+size 4877701896

model-00006-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3dd932c6e3ef4dbdfd5bdd27c32ae0b9cede00033c5c025c09127aafd14fcf0
+size 4877701896

model-00007-of-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea954cbeae92e996cff924211e5c53e13798ed570ced2ac212b417c951f55614
+size 4908582968

quant_w8a8.sh

@@ -0,0 +1,22 @@
+#!/bin/bash
+
+    # for DAMPENING_FRAC 0.01 0.1;
+    # do
+    #     for OBSERVER minmax mse;
+    #     do
+    #         for ACTORDER False group;
+    #         do
+
+# export DAMPENING_FRAC=0.01
+# export OBSERVER="minmax"
+
+export CUDA_VISIBLE_DEVICES=${1}
+export MDL=${2}
+export OBSERVER=${3} # minmax mse
+export DAMPENING_FRAC=${4} # 0.01 0.1
+
+for CALIB_SIZE in 128 512 1024;
+do
+    python w8a8.py --model_path ${MDL} --quant_path "output_dir_w8a8/${MDL}/calib${CALIB_SIZE}_eosFalse_damp${DAMPENING_FRAC}_obs${OBSERVER}" --calib_size ${CALIB_SIZE} --dampening_frac ${DAMPENING_FRAC} --observer ${OBSERVER}
+done
+

recipe.yaml

@@ -0,0 +1,7 @@
+DEFAULT_stage:
+  DEFAULT_modifiers:
+    GPTQModifier:
+      targets: [Linear]
+      dampening_frac: 0.1
+      ignore: [lm_head]
+      scheme: W8A8

special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_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|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

tokenizer_config.json

@@ -0,0 +1,208 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_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|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are a helpful and harmless assistant. You are Qwen developed by Alibaba. You should think step-by-step.' }}\n    {%- endif %}\n    {{- \"\\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>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\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\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are a helpful and harmless assistant. You are Qwen developed by Alibaba. You should think step-by-step.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 32768,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

w8a8.py

@@ -0,0 +1,123 @@
+from datasets import load_dataset
+from transformers import AutoTokenizer
+from llmcompressor.modifiers.quantization import GPTQModifier
+from llmcompressor.transformers import SparseAutoModelForCausalLM, oneshot, apply
+import argparse
+from compressed_tensors.quantization import QuantizationScheme, QuantizationArgs, QuantizationType, QuantizationStrategy
+
+
+parser = argparse.ArgumentParser()
+parser.add_argument('--model_path', type=str, default="/network/eldar/llama31_8b_gsm8k/sparse_transfer/sp2of4_2ep_lr7e-6_bs32_GradClip2_warmup20ba_noKD")
+parser.add_argument('--quant_path', type=str, default="output_dir/quant_test")
+
+parser.add_argument('--calib_size', type=int, default=256) # 32 64 128 256 512 1024 2048
+parser.add_argument('--dampening_frac', type=float, default=0.01) # 0.001 0.003 0.005 0.008 0.01 0.03 0.05 0.08 0.1 0.3
+parser.add_argument('--observer', type=str, default="minmax") # mse or minmax
+
+# ToDo: add args for SmoothQuant if needed
+args = parser.parse_args()
+
+print(f"[DEBUGGING ARGS] {args}")
+# TODO: to ablate on additionally is whether we want to add EOS token to calib data
+
+
+model = SparseAutoModelForCausalLM.from_pretrained(
+    args.model_path,
+    device_map="auto",
+    torch_dtype="auto",
+    use_cache=False,
+    trust_remote_code=True,
+)
+tokenizer = AutoTokenizer.from_pretrained(args.model_path)
+
+
+NUM_CALIBRATION_SAMPLES = args.calib_size
+DATASET_ID = "garage-bAInd/Open-Platypus"
+DATASET_SPLIT = "train"
+ds = load_dataset(DATASET_ID, split=DATASET_SPLIT)
+ds = ds.shuffle(seed=42).select(range(NUM_CALIBRATION_SAMPLES))
+
+def preprocess(example):
+    concat_txt = example["instruction"] + "\n" + example["output"]
+    return {"text": concat_txt}
+
+ds = ds.map(preprocess)
+print(f"================================================================================")
+print(f"[For debugging] Calibration data sample is:\n{repr(ds[0]['text'])}")
+print(f"================================================================================")
+
+def tokenize(sample):
+    return tokenizer(
+        sample["text"],
+        padding=False,
+        truncation=False,
+        add_special_tokens=True,
+    )
+
+
+ds = ds.map(tokenize, remove_columns=ds.column_names)
+print(f"================================================================================")
+print(f"[For debugging] Tokenized data sample is:\n{tokenizer.decode(ds[0]['input_ids'])}")
+print(f"================================================================================")
+
+# TODO: why do we need to fuck around the calibration data when it's given completely prepared
+# TODO: are recipe_args used to override params?
+
+#quant_scheme = QuantizationScheme(
+#    targets=["Linear"],
+#    weights=QuantizationArgs(
+#       num_bits=args.num_bits,
+#       type=QuantizationType.INT,
+#       symmetric=True,
+#       group_size=128,
+#       strategy=QuantizationStrategy.GROUP,
+#       observer=args.observer,
+#       actorder=args.actorder
+#   ),
+#   input_activations=None,
+#   output_activations=None,
+#)
+
+recipe = [
+    GPTQModifier(
+        targets=["Linear"],
+        ignore=["lm_head"],
+        scheme="W8A8",
+        dampening_frac=args.dampening_frac,
+        observer=args.observer,
+        #config_groups={"group_0": quant_scheme},
+    )
+]
+oneshot(
+    model=model,
+    dataset=ds,
+    # recipe="w4a16_recipe.yaml",
+    recipe=recipe,
+    # recipe_args={
+    #     "quant_stage.quant_modifiers.GPTQModifier.dampening_frac": 100,
+    #     "quant_stage.quant_modifiers.GPTQModifier.sequential_update": args.sequential_update,
+    #     "quant_stage.quant_modifiers.GPTQModifier.config_groups.group_0.weights.observer": args.observer,
+    #     "quant_stage.quant_modifiers.GPTQModifier.config_groups.group_0.weights.actorder": args.actorder,
+    # },
+    num_calibration_samples=args.calib_size,
+    max_seq_length=8192,
+)
+
+
+# apply(
+#     recipe=self.recipe,
+#     # recipe_stage=None,
+#     recipe_args=self.recipe_args,
+#     model=self.model,
+#     calib_data=calibration_data, # dataloader
+#     start=-1,
+#     copy_data=False,
+#     accelerator=self.accelerator, # some accelerator object
+#     min_tokens_per_module=None
+# )
+
+# Save to disk compressed.
+SAVE_DIR = args.quant_path
+model.save_pretrained(SAVE_DIR, save_compressed=True)
+tokenizer.save_pretrained(SAVE_DIR)
+