Create README.md

README.md

@@ -0,0 +1,263 @@
+---
+language:
+- en
+pipeline_tag: text-generation
+license: apache-2.0
+license_link: https://www.apache.org/licenses/LICENSE-2.0
+---
+
+# Qwen2-0.5B-Instruct-quantized.w8a8
+
+## Model Overview
+- **Model Architecture:** Qwen2
+  - **Input:** Text
+  - **Output:** Text
+- **Model Optimizations:**
+  - **Activation quantization:** INT8
+  - **Weight quantization:** INT8
+- **Intended Use Cases:** Intended for commercial and research use in English. Similarly to [Qwen2-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2-0.5B-Instruct), this models is intended for assistant-like chat.
+- **Out-of-scope:** Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in languages other than English.
+- **Release Date:** 7/15/2024
+- **Version:** 1.0
+- **License(s):** [Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0)
+- **Model Developers:** Neural Magic
+
+Quantized version of [Qwen2-72B-Instruct](https://huggingface.co/Qwen/Qwen2-72B-Instruct).
+It achieves an average score of 80.32 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 80.09.
+
+### Model Optimizations
+
+This model was obtained by quantizing the weights of [Qwen2-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2-0.5B-Instruct) to INT8 data type.
+This optimization reduces the number of bits used to represent weights and activations from 16 to 8, reducing GPU memory requirements (by approximately 50%) and increasing matrix-multiply compute throughput (by approximately 2x).
+Weight quantization also reduces disk size requirements by approximately 50%.
+
+Only weights and activations of the linear operators within transformers blocks are quantized.
+Weights are quantized with a symmetric static per-channel scheme, where a fixed linear scaling factor is applied between INT8 and floating point representations for each output channel dimension.
+Activations are quantized with a symmetric dynamic per-token scheme, computing a linear scaling factor at runtime for each token between INT8 and floating point representations.
+The [GPTQ](https://arxiv.org/abs/2210.17323) algorithm is applied for quantization, as implemented in the [llm-compressor](https://github.com/vllm-project/llm-compressor) library.
+GPTQ used a 1% damping factor and 256 sequences of 8,192 random tokens.
+
+
+## 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 (using 2 GPUs).
+
+```python
+from vllm import LLM, SamplingParams
+from transformers import AutoTokenizer
+
+model_id = "neuralmagic/Qwen2-72B-Instruct-quantized.w8a8"
+number_gpus = 2
+
+sampling_params = SamplingParams(temperature=0.7, top_p=0.8, max_tokens=256)
+
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+
+messages = [
+    {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
+    {"role": "user", "content": "Who are you?"},
+]
+
+prompts = tokenizer.apply_chat_template(messages, add_generation_prompt=True, tokenize=False)
+
+llm = LLM(model=model_id, tensor_parallel_size=number_gpus)
+
+outputs = llm.generate(prompts, sampling_params)
+
+generated_text = outputs[0].outputs[0].text
+print(generated_text)
+```
+
+vLLM aslo supports OpenAI-compatible serving. See the [documentation](https://docs.vllm.ai/en/latest/) for more details.
+
+### Use with transformers
+
+The following example contemplates how the model can be deployed in Transformers using the `generate()` function.
+
+```python
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+model_id = "neuralmagic/Qwen2-72B-Instruct-quantized.w8a8"
+
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+model = AutoModelForCausalLM.from_pretrained(
+    model_id,
+    torch_dtype="auto",
+    device_map="auto",
+)
+
+messages = [
+    {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
+    {"role": "user", "content": "Who are you?"},
+]
+
+input_ids = tokenizer.apply_chat_template(
+    messages,
+    add_generation_prompt=True,
+    return_tensors="pt"
+).to(model.device)
+
+terminators = [
+    tokenizer.eos_token_id,
+    tokenizer.convert_tokens_to_ids("<|eot_id|>")
+]
+
+outputs = model.generate(
+    input_ids,
+    max_new_tokens=256,
+    eos_token_id=terminators,
+    do_sample=True,
+    temperature=0.7,
+    top_p=0.8,
+)
+response = outputs[0][input_ids.shape[-1]:]
+print(tokenizer.decode(response, skip_special_tokens=True))
+```
+
+## Creation
+
+This model was created by using the [llm-compressor](https://github.com/vllm-project/llm-compressor) library as presented in the code snipet below.
+
+```python
+from transformers import AutoTokenizer
+from datasets import Dataset
+from llmcompressor.transformers import SparseAutoModelForCausalLM, oneshot
+from llmcompressor.modifiers.quantization import GPTQModifier
+import random
+
+model_id = "Qwen/Qwen2-72B-Instruct"
+
+num_samples = 256
+max_seq_len = 8192
+
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+
+max_token_id = len(tokenizer.get_vocab()) - 1
+input_ids = [[random.randint(0, max_token_id) for _ in range(max_seq_len)] for _ in range(num_samples)]
+attention_mask = num_samples * [max_seq_len * [1]]
+ds = Dataset.from_dict({"input_ids": input_ids, "attention_mask": attention_mask})
+
+recipe = GPTQModifier(
+  targets="Linear",
+  scheme="W8A8",
+  ignore=["lm_head"],
+  dampening_frac=0.01,
+)
+
+model = SparseAutoModelForCausalLM.from_pretrained(
+  model_id,
+  device_map="auto",
+  trust_remote_code=True,
+)
+
+oneshot(
+  model=model,
+  dataset=ds,
+  recipe=recipe,
+  max_seq_length=max_seq_len,
+  num_calibration_samples=num_samples,
+)
+
+model.save_pretrained("Qwen2-72B-Instruct-quantized.w8a18)
+```
+
+
+
+## Evaluation
+
+The model was evaluated on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) leaderboard tasks (version 1) with the [lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness/tree/383bbd54bc621086e05aa1b030d8d4d5635b25e6) (commit 383bbd54bc621086e05aa1b030d8d4d5635b25e6) and the [vLLM](https://docs.vllm.ai/en/stable/) engine, using the following command (using 2 GPUs):
+```
+lm_eval \
+  --model vllm \
+  --model_args pretrained="neuralmagic/Qwen2-72B-Instruct-quantized.w8a8",dtype=auto,gpu_memory_utilization=0.4,add_bos_token=True,max_model_len=4096,tensor_parallel_size=2 \
+  --tasks openllm \
+  --batch_size auto
+```
+
+### Accuracy
+
+#### Open LLM Leaderboard evaluation scores
+<table>
+  <tr>
+   <td><strong>Benchmark</strong>
+   </td>
+   <td><strong>Qwen2-72B-Instruct</strong>
+   </td>
+   <td><strong>Qwen2-72B-Instruct-quantized.w8a8 (this model)</strong>
+   </td>
+   <td><strong>Recovery</strong>
+   </td>
+  </tr>
+  <tr>
+   <td>MMLU (5-shot)
+   </td>
+   <td>83.97
+   </td>
+   <td>83.78
+   </td>
+   <td>99.8%
+   </td>
+  </tr>
+  <tr>
+   <td>ARC Challenge (25-shot)
+   </td>
+   <td>71.59
+   </td>
+   <td>72.01
+   </td>
+   <td>100.6%
+   </td>
+  </tr>
+  <tr>
+   <td>GSM-8K (5-shot, strict-match)
+   </td>
+   <td>88.25
+   </td>
+   <td>88.93
+   </td>
+   <td>100.8%
+   </td>
+  </tr>
+  <tr>
+   <td>Hellaswag (10-shot)
+   </td>
+   <td>86.94
+   </td>
+   <td>87.18
+   </td>
+   <td>100.3%
+   </td>
+  </tr>
+  <tr>
+   <td>Winogrande (5-shot)
+   </td>
+   <td>82.79
+   </td>
+   <td>83.35
+   </td>
+   <td>100.7%
+   </td>
+  </tr>
+  <tr>
+   <td>TruthfulQA (0-shot)
+   </td>
+   <td>66.98
+   </td>
+   <td>66.65
+   </td>
+   <td>99.5%
+   </td>
+  </tr>
+  <tr>
+   <td><strong>Average</strong>
+   </td>
+   <td><strong>80.09</strong>
+   </td>
+   <td><strong>80.32</strong>
+   </td>
+   <td><strong>100.3%</strong>
+   </td>
+  </tr>
+</table>