RedHatAI
/

Qwen2.5-14B-FP8-dynamic

@@ -1,168 +1,180 @@
----
-license: apache-2.0
-license_link: https://huggingface.co/Qwen/Qwen2.5-14B/blob/main/LICENSE
-language:
-- en
-pipeline_tag: text-generation
-base_model: Qwen/Qwen2.5-14B
-tags:
-- chat
-- neuralmagic
-- llmcompressor
----
-# Qwen2.5-14B-FP8-dynamic
-## 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 multiple languages. Similarly to [Qwen2.5-14B](https://huggingface.co/Qwen/Qwen2.5-14B), 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).
-- **Release Date:** 11/27/2024
-- **Version:** 1.0
-- **License(s):** [apache-2.0](https://huggingface.co/Qwen/Qwen2.5-14B/blob/main/LICENSE)
-- **Model Developers:** Neural Magic
-Quantized version of [Qwen2.5-14B](https://huggingface.co/Qwen/Qwen2.5-14B).
-It achieves an average score of 75.44 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 75.66.
-### Model Optimizations
-This model was obtained by quantizing the weights of [Qwen2.5-14B](https://huggingface.co/Qwen/Qwen2.5-14B) 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.
-## Deployment
-This model can be deployed efficiently using the [vLLM](https://docs.vllm.ai/en/latest/) backend, as shown in the example below.
-```python
-from vllm import LLM, SamplingParams
-from transformers import AutoTokenizer
-model_id = "neuralmagic/Qwen2.5-14B-FP8-dynamic"
-number_gpus = 1
-max_model_len = 8192
-sampling_params = SamplingParams(temperature=0.7, top_p=0.8, max_tokens=256)
-tokenizer = AutoTokenizer.from_pretrained(model_id)
-prompt = "Give me a short introduction to large language model."
-llm = LLM(model=model_id, tensor_parallel_size=number_gpus, max_model_len=max_model_len)
-outputs = llm.generate(prompt, 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.
-## 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:
-```
-lm_eval \
-  --model vllm \
-  --model_args pretrained="neuralmagic/Qwen2.5-14B-FP8-dynamic",dtype=auto,gpu_memory_utilization=0.9,add_bos_token=True,max_model_len=4096,enable_chunk_prefill=True,tensor_parallel_size=1 \
-  --tasks openllm \
-  --batch_size auto
-```
-### Accuracy
-#### Open LLM Leaderboard evaluation scores
-<table>
-  <tr>
-   <td><strong>Benchmark</strong>
-   </td>
-   <td><strong>Qwen2.5-14B</strong>
-   </td>
-   <td><strong>Qwen2.5-14B-FP8-dynamic (this model)</strong>
-   </td>
-   <td><strong>Recovery</strong>
-   </td>
-  </tr>
-  <tr>
-   <td>MMLU (5-shot)
-   </td>
-   <td>79.71
-   </td>
-   <td>79.48
-   </td>
-   <td>99.7%
-   </td>
-  </tr>
-  <tr>
-   <td>ARC Challenge (25-shot)
-   </td>
-   <td>65.70
-   </td>
-   <td>65.19
-   </td>
-   <td>99.2%
-   </td>
-  </tr>
-  <tr>
-   <td>GSM-8K (5-shot, strict-match)
-   </td>
-   <td>84.46
-   </td>
-   <td>83.62
-   </td>
-   <td>99.0%
-   </td>
-  </tr>
-  <tr>
-   <td>Hellaswag (10-shot)
-   </td>
-   <td>84.28
-   </td>
-   <td>84.28
-   </td>
-   <td>100.0%
-   </td>
-  </tr>
-  <tr>
-   <td>Winogrande (5-shot)
-   </td>
-   <td>81.37
-   </td>
-   <td>81.77
-   </td>
-   <td>100.5%
-   </td>
-  </tr>
-  <tr>
-   <td>TruthfulQA (0-shot, mc2)
-   </td>
-   <td>58.46
-   </td>
-   <td>58.30
-   </td>
-   <td>99.7%
-   </td>
-  </tr>
-  <tr>
-   <td><strong>Average</strong>
-   </td>
-   <td><strong>75.66</strong>
-   </td>
-   <td><strong>75.44</strong>
-   </td>
-   <td><strong>99.7%</strong>
-   </td>
-  </tr>
-</table>

+---
+license: apache-2.0
+license_link: https://huggingface.co/Qwen/Qwen2.5-14B/blob/main/LICENSE
+language:
+- zho
+- eng
+- fra
+- spa
+- por
+- deu
+- ita
+- rus
+- jpn
+- kor
+- vie
+- tha
+- ara
+pipeline_tag: text-generation
+base_model: Qwen/Qwen2.5-14B
+tags:
+- chat
+- neuralmagic
+- llmcompressor
+---
+# Qwen2.5-14B-FP8-dynamic
+## 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 multiple languages. Similarly to [Qwen2.5-14B](https://huggingface.co/Qwen/Qwen2.5-14B), 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).
+- **Release Date:** 11/27/2024
+- **Version:** 1.0
+- **License(s):** [apache-2.0](https://huggingface.co/Qwen/Qwen2.5-14B/blob/main/LICENSE)
+- **Model Developers:** Neural Magic
+Quantized version of [Qwen2.5-14B](https://huggingface.co/Qwen/Qwen2.5-14B).
+It achieves an average score of 75.44 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 75.66.
+### Model Optimizations
+This model was obtained by quantizing the weights of [Qwen2.5-14B](https://huggingface.co/Qwen/Qwen2.5-14B) 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.
+## Deployment
+This model can be deployed efficiently using the [vLLM](https://docs.vllm.ai/en/latest/) backend, as shown in the example below.
+```python
+from vllm import LLM, SamplingParams
+from transformers import AutoTokenizer
+model_id = "neuralmagic/Qwen2.5-14B-FP8-dynamic"
+number_gpus = 1
+max_model_len = 8192
+sampling_params = SamplingParams(temperature=0.7, top_p=0.8, max_tokens=256)
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+prompt = "Give me a short introduction to large language model."
+llm = LLM(model=model_id, tensor_parallel_size=number_gpus, max_model_len=max_model_len)
+outputs = llm.generate(prompt, 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.
+## 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:
+```
+lm_eval \
+  --model vllm \
+  --model_args pretrained="neuralmagic/Qwen2.5-14B-FP8-dynamic",dtype=auto,gpu_memory_utilization=0.9,add_bos_token=True,max_model_len=4096,enable_chunk_prefill=True,tensor_parallel_size=1 \
+  --tasks openllm \
+  --batch_size auto
+```
+### Accuracy
+#### Open LLM Leaderboard evaluation scores
+<table>
+  <tr>
+   <td><strong>Benchmark</strong>
+   </td>
+   <td><strong>Qwen2.5-14B</strong>
+   </td>
+   <td><strong>Qwen2.5-14B-FP8-dynamic (this model)</strong>
+   </td>
+   <td><strong>Recovery</strong>
+   </td>
+  </tr>
+  <tr>
+   <td>MMLU (5-shot)
+   </td>
+   <td>79.71
+   </td>
+   <td>79.48
+   </td>
+   <td>99.7%
+   </td>
+  </tr>
+  <tr>
+   <td>ARC Challenge (25-shot)
+   </td>
+   <td>65.70
+   </td>
+   <td>65.19
+   </td>
+   <td>99.2%
+   </td>
+  </tr>
+  <tr>
+   <td>GSM-8K (5-shot, strict-match)
+   </td>
+   <td>84.46
+   </td>
+   <td>83.62
+   </td>
+   <td>99.0%
+   </td>
+  </tr>
+  <tr>
+   <td>Hellaswag (10-shot)
+   </td>
+   <td>84.28
+   </td>
+   <td>84.28
+   </td>
+   <td>100.0%
+   </td>
+  </tr>
+  <tr>
+   <td>Winogrande (5-shot)
+   </td>
+   <td>81.37
+   </td>
+   <td>81.77
+   </td>
+   <td>100.5%
+   </td>
+  </tr>
+  <tr>
+   <td>TruthfulQA (0-shot, mc2)
+   </td>
+   <td>58.46
+   </td>
+   <td>58.30
+   </td>
+   <td>99.7%
+   </td>
+  </tr>
+  <tr>
+   <td><strong>Average</strong>
+   </td>
+   <td><strong>75.66</strong>
+   </td>
+   <td><strong>75.44</strong>
+   </td>
+   <td><strong>99.7%</strong>
+   </td>
+  </tr>
+</table>