|
--- |
|
license: mit |
|
tags: |
|
- deepseek |
|
- fp8 |
|
- vllm |
|
base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B |
|
library_name: transformers |
|
--- |
|
|
|
# DeepSeek-R1-Distill-Qwen-1.5B-FP8-dynamic |
|
|
|
## Model Overview |
|
- **Model Architecture:** Qwen2ForCausalLM |
|
- **Input:** Text |
|
- **Output:** Text |
|
- **Model Optimizations:** |
|
- **Weight quantization:** FP8 |
|
- **Activation quantization:** FP8 |
|
- **Release Date:** 2/5/2025 |
|
- **Version:** 1.0 |
|
- **Model Developers:** Neural Magic |
|
|
|
Quantized version of [DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B). |
|
|
|
|
|
### Model Optimizations |
|
|
|
This model was obtained by quantizing the weights and activations of [DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B) to FP8 data type. |
|
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. |
|
Weights are quantized using a symmetric per-channel scheme, whereas quantizations are quantized using a symmetric per-token scheme. |
|
[LLM Compressor](https://github.com/vllm-project/llm-compressor) is used for quantization. |
|
|
|
|
|
## 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 |
|
|
|
number_gpus = 1 |
|
model_name = "neuralmagic/DeepSeek-R1-Distill-Qwen-1.5B-dynamic" |
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model_name) |
|
sampling_params = SamplingParams(temperature=0.6, max_tokens=256, stop_token_ids=[tokenizer.eos_token_id]) |
|
llm = LLM(model=model_name, tensor_parallel_size=number_gpus, trust_remote_code=True) |
|
|
|
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. |
|
|
|
|
|
```python |
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
from llmcompressor.modifiers.quantization import QuantizationModifier |
|
from llmcompressor.transformers import oneshot |
|
import os |
|
|
|
# Load model |
|
model_stub = "deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B" |
|
model_name = model_stub.split("/")[-1] |
|
|
|
model = AutoModelForCausalLM.from_pretrained( |
|
model_stub, |
|
torch_dtype="auto", |
|
) |
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model_stub) |
|
|
|
# Configure the quantization algorithm and scheme |
|
recipe = QuantizationModifier( |
|
targets="Linear", |
|
scheme="FP8_DYNAMIC", |
|
ignore=["lm_head"], |
|
) |
|
|
|
# Apply quantization |
|
oneshot( |
|
model=model, |
|
recipe=recipe, |
|
) |
|
|
|
# Save to disk in compressed-tensors format |
|
save_path = model_name + "-FP8-dynamic |
|
model.save_pretrained(save_path) |
|
tokenizer.save_pretrained(save_path) |
|
print(f"Model and tokenizer saved to: {save_path}") |
|
``` |
|
|
|
## 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/DeepSeek-R1-Distill-Qwen-1.5B-FP8-dynamic",dtype=auto,max_model_len=4096,tensor_parallel_size=1,enable_chunked_prefill=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/DeepSeek-R1-Distill-Qwen-1.5B-FP8-dynamic",dtype=auto,max_model_len=4096,tensor_parallel_size=1,enable_chunked_prefill=True \ |
|
--apply_chat_template \ |
|
--fewshot_as_multiturn \ |
|
--tasks leaderboard \ |
|
--write_out \ |
|
--batch_size auto \ |
|
--output_path output_dir \ |
|
--show_config |
|
``` |
|
|
|
### Accuracy |
|
|
|
<table> |
|
<thead> |
|
<tr> |
|
<th>Category</th> |
|
<th>Metric</th> |
|
<th>deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B</th> |
|
<th>neuralmagic/DeepSeek-R1-Distill-Qwen-1.5B-FP8-dynamic</th> |
|
<th>Recovery</th> |
|
</tr> |
|
</thead> |
|
<tbody> |
|
<tr> |
|
<td rowspan="7"><b>OpenLLM V1</b></td> |
|
<td>ARC-Challenge (Acc-Norm, 25-shot)</td> |
|
<td>37.02</td> |
|
<td>37.71</td> |
|
<td>101.4%</td> |
|
</tr> |
|
<tr> |
|
<td>GSM8K (Strict-Match, 5-shot)</td> |
|
<td>69.98</td> |
|
<td>68.99</td> |
|
<td>98.6%</td> |
|
</tr> |
|
<tr> |
|
<td>HellaSwag (Acc-Norm, 10-shot)</td> |
|
<td>43.86</td> |
|
<td>43.61</td> |
|
<td>99.4%</td> |
|
</tr> |
|
<tr> |
|
<td>MMLU (Acc, 5-shot)</td> |
|
<td>37.38</td> |
|
<td>37.22</td> |
|
<td>99.6%</td> |
|
</tr> |
|
<tr> |
|
<td>TruthfulQA (MC2, 0-shot)</td> |
|
<td>45.21</td> |
|
<td>44.77</td> |
|
<td>99.0%</td> |
|
</tr> |
|
<tr> |
|
<td>Winogrande (Acc, 5-shot)</td> |
|
<td>54.30</td> |
|
<td>54.62</td> |
|
<td>100.6%</td> |
|
</tr> |
|
<tr> |
|
<td><b>Average Score</b></td> |
|
<td><b>47.99</b></td> |
|
<td><b>47.82</b></td> |
|
<td><b>99.7%</b></td> |
|
</tr> |
|
<tr> |
|
<td rowspan="7"><b>OpenLLM V2</b></td> |
|
<td>IFEval (Inst Level Strict Acc, 0-shot)</td> |
|
<td>34.37</td> |
|
<td>34.91</td> |
|
<td>101.6%</td> |
|
</tr> |
|
<tr> |
|
<td>BBH (Acc-Norm, 3-shot)</td> |
|
<td>34.44</td> |
|
<td>34.40</td> |
|
<td>99.9%</td> |
|
</tr> |
|
<tr> |
|
<td>Math-Hard (Exact-Match, 4-shot)</td> |
|
<td>0.00</td> |
|
<td>0.00</td> |
|
<td>---</td> |
|
</tr> |
|
<tr> |
|
<td>GPQA (Acc-Norm, 0-shot)</td> |
|
<td>24.67</td> |
|
<td>25.16</td> |
|
<td>102.0%</td> |
|
</tr> |
|
<tr> |
|
<td>MUSR (Acc-Norm, 0-shot)</td> |
|
<td>35.82</td> |
|
<td>36.61</td> |
|
<td>102.2%</td> |
|
</tr> |
|
<tr> |
|
<td>MMLU-Pro (Acc, 5-shot)</td> |
|
<td>11.80</td> |
|
<td>11.69</td> |
|
<td>99.1%</td> |
|
</tr> |
|
<tr> |
|
<td><b>Average Score</b></td> |
|
<td><b>23.52</b></td> |
|
<td><b>23.79</b></td> |
|
<td><b>101.2%</b></td> |
|
</tr> |
|
<tr> |
|
<td rowspan="4"><b>Coding</b></td> |
|
<td>HumanEval (pass@1)</td> |
|
<td>37.90</td> |
|
<td>36.40</td> |
|
<td><b>96.0%</b></td> |
|
</tr> |
|
<tr> |
|
<td>HumanEval (pass@10)</td> |
|
<td>61.30</td> |
|
<td>61.30</td> |
|
<td>100.0%</td> |
|
</tr> |
|
<tr> |
|
<td>HumanEval+ (pass@10)</td> |
|
<td>33.00</td> |
|
<td>32.60</td> |
|
<td>98.8%</td> |
|
</tr> |
|
<tr> |
|
<td>HumanEval+ (pass@10)</td> |
|
<td>55.90</td> |
|
<td>56.30</td> |
|
<td>100.7%</td> |
|
</tr> |
|
</tbody> |
|
</table> |
|
|
