File size: 5,647 Bytes
544606b 2314c91 544606b 2314c91 4200ed6 2314c91 4200ed6 2314c91 4200ed6 2314c91 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 |
---
datasets: wikitext
license: apache-2.0
license_link: https://llama.meta.com/llama3/license/
---
This is a quantized model of [Llama-3 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct) using GPTQ developed by [IST Austria](https://ist.ac.at/en/research/alistarh-group/)
using the following configuration:
- 4bit (8bit will follow)
- Act order: True
- Group size: 128
- Seq. length: 4096
## Usage
Install **vLLM** and
run the [server](https://docs.vllm.ai/en/latest/serving/openai_compatible_server.html#openai-compatible-server):
```
python -m vllm.entrypoints.openai.api_server --model cortecs/Meta-Llama-3-70B-Instruct-GPTQ
```
Access the model:
```
curl http://localhost:8000/v1/completions
-H "Content-Type: application/json"
-d '{
"model": "cortecs/Meta-Llama-3-70B-Instruct-GPTQ",
"prompt": "<|begin_of_text|><|start_header_id|>user<|end_header_id|>
Tell me a joke<|eot_id|><|start_header_id|>assistant<|end_header_id|>"
}'
```
## Evaluations
| __English__ | __Llama-3 70B Instruct__ | __Llama 3 70B GPTQ__ | __Llama-3 8B Instruct__ |
|:--------------|:---------------------------|:-----------------------|:--------------------------|
| Avg. | 76.19 | 75.14 | 66.97 |
| ARC | 71.6 | 70.7 | 62.5 |
| Hellaswag | 77.3 | 76.4 | 70.3 |
| MMLU | 79.66 | 78.33 | 68.11 |
| | | | |
| __French__ | __Llama-3 70B Instruct__ | __Llama 3 70B GPTQ__ | __Llama-3 8B Instruct__ |
| Avg. | 70.97 | 70.27 | 57.73 |
| ARC_fr | 65.0 | 64.7 | 53.3 |
| Hellaswag_fr | 72.4 | 71.4 | 61.7 |
| MMLU_fr | 75.5 | 74.7 | 58.2 |
| | | | |
| __German__ | __Llama-3 70B Instruct__ | __Llama 3 70B GPTQ__ | __Llama-3 8B Instruct__ |
| Avg. | 68.43 | 66.93 | 53.47 |
| ARC_de | 64.2 | 62.6 | 49.1 |
| Hellaswag_de | 67.8 | 66.7 | 55.0 |
| MMLU_de | 73.3 | 71.5 | 56.3 |
| | | | |
| __Italian__ | __Llama-3 70B Instruct__ | __Llama 3 70B GPTQ__ | __Llama-3 8B Instruct__ |
| Avg. | 70.17 | 68.63 | 56.73 |
| ARC_it | 64.0 | 62.1 | 51.6 |
| Hellaswag_it | 72.6 | 71.0 | 61.3 |
| MMLU_it | 73.9 | 72.8 | 57.3 |
| | | | |
| __Safety__ | __Llama-3 70B Instruct__ | __Llama 3 70B GPTQ__ | __Llama-3 8B Instruct__ |
| Avg. | 64.28 | 63.64 | 61.42 |
| RealToxicityPrompts | 97.9 | 98.1 | 97.2 |
| TruthfulQA | 61.91 | 59.91 | 51.65 |
| CrowS | 33.04 | 32.92 | 35.42 |
| | | | |
| __Spanish__ | __Llama-3 70B Instruct__ | __Llama 3 70B GPTQ__ | __Llama-3 8B Instruct__ |
| Avg. | 72.5 | 71.3 | 59 |
| ARC_es | 66.7 | 65.7 | 54.1 |
| Hellaswag_es | 75.8 | 74 | 63.8 |
| MMLU_es | 75 | 74.2 | 59.1 |
Take with caution. We did not check for data contamination.
Evaluation was done using [Eval. Harness](https://github.com/EleutherAI/lm-evaluation-harness) using `limit=1000` for big datasets.
## Performance
| __Llama-3 70B Instruct__ | __requests/s__ | __tokens/s__ |
|:---------------------------|:-----------------|:---------------|
| NVIDIA L40Sx4 | 2.38 | 1135.41 |
| | | |
| __Llama 3 70B GPTQ__ | __requests/s__ | __tokens/s__ |
| NVIDIA L40Sx2 | 2.0 | 951.28 |
| | | |
| __Llama-3 8B Instruct__ | __requests/s__ | __tokens/s__ |
| NVIDIA L40Sx1 | 11.64 | 5548.63 |
| NVIDIA L4x1 | 2.76 | 1315.25 |
| NVIDIA L4x2 | 4.79 | 2283.53 |
Performance was measured on [cortecs.ai](https://cortecs.ai). |