alexmarques's picture
Update README.md
f65789c verified
---
tags:
- fp8
- vllm
language:
- en
- de
- fr
- it
- pt
- hi
- es
- th
pipeline_tag: text-generation
license: llama3.2
base_model: meta-llama/Llama-3.2-1B-Instruct
---
# Llama-3.2-1B-Instruct-FP8-dynamic
## Model Overview
- **Model Architecture:** Meta-Llama-3.2
- **Input:** Text
- **Output:** Text
- **Model Optimizations:**
- **Weight quantization:** FP8
- **Activation quantization:** FP8
- **Intended Use Cases:** Intended for commercial and research use in multiple languages. Similarly to [Llama-3.2-1B-Instruct](https://huggingface.co/meta-llama/Llama-3.2-1B-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:** 9/25/2024
- **Version:** 1.0
- **License(s):** [llama3.2](https://huggingface.co/meta-llama/Llama-3.2-1B/blob/main/LICENSE)
- **Model Developers:** Neural Magic
Quantized version of [Llama-3.2-1B-Instruct](https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct).
It achieves an average score of 50.88 on a subset of task from the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 51.70.
### Model Optimizations
This model was obtained by quantizing the weights and activations of [Llama-3.2-1B-Instruct](https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct) to FP8 data type, ready for inference with vLLM built from source.
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. Symmetric per-channel quantization is applied, in which a linear scaling per output dimension maps the FP8 representations of the quantized weights and activations. Activations are also quantized on a per-token dynamic basis.
[LLM Compressor](https://github.com/vllm-project/llm-compressor) is used for quantization.
## 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 vllm import LLM, SamplingParams
from transformers import AutoTokenizer
model_id = "neuralmagic/Llama-3.2-1B-Instruct-FP8-dynamic"
sampling_params = SamplingParams(temperature=0.6, top_p=0.9, 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, tokenize=False)
llm = LLM(model=model_id)
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.
## Creation
This model was created by applying [LLM Compressor](https://github.com/vllm-project/llm-compressor/blob/sa/big_model_support/examples/big_model_offloading/big_model_w8a8_calibrate.py), as presented in the code snipet below.
```python
import torch
from transformers import AutoTokenizer
from llmcompressor.transformers import SparseAutoModelForCausalLM, oneshot
from llmcompressor.transformers.compression.helpers import ( # noqa
calculate_offload_device_map,
custom_offload_device_map,
)
recipe = """
quant_stage:
quant_modifiers:
QuantizationModifier:
ignore: ["lm_head"]
config_groups:
group_0:
weights:
num_bits: 8
type: float
strategy: channel
dynamic: false
symmetric: true
input_activations:
num_bits: 8
type: float
strategy: token
dynamic: true
symmetric: true
targets: ["Linear"]
"""
model_stub = "meta-llama/Llama-3.2-1B-Instruct"
model_name = model_stub.split("/")[-1]
device_map = calculate_offload_device_map(
model_stub, reserve_for_hessians=False, num_gpus=1, torch_dtype="auto"
)
model = SparseAutoModelForCausalLM.from_pretrained(
model_stub, torch_dtype="auto", device_map=device_map
)
output_dir = f"./{model_name}-FP8-dynamic"
oneshot(
model=model,
recipe=recipe,
output_dir=output_dir,
save_compressed=True,
tokenizer=AutoTokenizer.from_pretrained(model_stub),
)
```
## Evaluation
The model was evaluated on MMLU, ARC-Challenge, GSM-8K, and Winogrande.
Evaluation was conducted using the Neural Magic fork of [lm-evaluation-harness](https://github.com/neuralmagic/lm-evaluation-harness/tree/llama_3.1_instruct) (branch llama_3.1_instruct) and the [vLLM](https://docs.vllm.ai/en/stable/) engine.
This version of the lm-evaluation-harness includes versions of ARC-Challenge, GSM-8K, MMLU, and MMLU-cot that match the prompting style of [Meta-Llama-3.1-Instruct-evals](https://huggingface.co/datasets/meta-llama/Meta-Llama-3.1-8B-Instruct-evals).
### Accuracy
#### Open LLM Leaderboard evaluation scores
<table>
<tr>
<td><strong>Benchmark</strong>
</td>
<td><strong>Llama-3.2-1B-Instruct </strong>
</td>
<td><strong>Llama-3.2-1B-Instruct-FP8-dynamic (this model)</strong>
</td>
<td><strong>Recovery</strong>
</td>
</tr>
<tr>
<td>MMLU (5-shot)
</td>
<td>47.66
</td>
<td>47.55
</td>
<td>99.8%
</td>
</tr>
<tr>
<td>MMLU-cot (0-shot)
</td>
<td>47.10
</td>
<td>46.79
</td>
<td>99.3%
</td>
</tr>
<tr>
<td>ARC Challenge (0-shot)
</td>
<td>58.36
</td>
<td>57.25
</td>
<td>98.1%
</td>
</tr>
<tr>
<td>GSM-8K-cot (8-shot, strict-match)
</td>
<td>45.72
</td>
<td>45.94
</td>
<td>100.5%
</td>
</tr>
<tr>
<td>Winogrande (5-shot)
</td>
<td>62.27
</td>
<td>61.40
</td>
<td>98.6%
</td>
</tr>
<tr>
<td>Hellaswag (10-shot)
</td>
<td>61.01
</td>
<td>60.95
</td>
<td>99.9%
</td>
</tr>
<tr>
<td>TruthfulQA (0-shot, mc2)
</td>
<td>43.52
</td>
<td>44.23
</td>
<td>101.6%
</td>
</tr>
<tr>
<td><strong>Average</strong>
</td>
<td><strong>52.24</strong>
</td>
<td><strong>52.02</strong>
</td>
<td><strong>99.7%</strong>
</td>
</tr>
</table>
### Reproduction
The results were obtained using the following commands:
#### MMLU
```
lm_eval \
--model vllm \
--model_args pretrained="neuralmagic/Llama-3.2-1B-Instruct-FP8-dynamic",dtype=auto,max_model_len=3850,max_gen_toks=10,tensor_parallel_size=1 \
--tasks mmlu_llama_3.1_instruct \
--fewshot_as_multiturn \
--apply_chat_template \
--num_fewshot 5 \
--batch_size auto
```
#### MMLU-CoT
```
lm_eval \
--model vllm \
--model_args pretrained="neuralmagic/Llama-3.2-1B-Instruct-FP8-dynamic",dtype=auto,max_model_len=4064,max_gen_toks=1024,tensor_parallel_size=1 \
--tasks mmlu_cot_0shot_llama_3.1_instruct \
--apply_chat_template \
--num_fewshot 0 \
--batch_size auto
```
#### ARC-Challenge
```
lm_eval \
--model vllm \
--model_args pretrained="neuralmagic/Llama-3.2-1B-Instruct-FP8-dynamic",dtype=auto,max_model_len=3940,max_gen_toks=100,tensor_parallel_size=1 \
--tasks arc_challenge_llama_3.1_instruct \
--apply_chat_template \
--num_fewshot 0 \
--batch_size auto
```
#### GSM-8K
```
lm_eval \
--model vllm \
--model_args pretrained="neuralmagic/Llama-3.2-1B-Instruct-FP8-dynamic",dtype=auto,max_model_len=4096,max_gen_toks=1024,tensor_parallel_size=1 \
--tasks gsm8k_cot_llama_3.1_instruct \
--fewshot_as_multiturn \
--apply_chat_template \
--num_fewshot 8 \
--batch_size auto
```
#### Hellaswag
```
lm_eval \
--model vllm \
--model_args pretrained="neuralmagic/Llama-3.2-1B-Instruct-FP8-dynamic",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=1 \
--tasks hellaswag \
--num_fewshot 10 \
--batch_size auto
```
#### Winogrande
```
lm_eval \
--model vllm \
--model_args pretrained="neuralmagic/Llama-3.2-1B-Instruct-FP8-dynamic",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=1 \
--tasks winogrande \
--num_fewshot 5 \
--batch_size auto
```
#### TruthfulQA
```
lm_eval \
--model vllm \
--model_args pretrained="neuralmagic/Llama-3.2-1B-Instruct-FP8-dynamic",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=1 \
--tasks truthfulqa \
--num_fewshot 0 \
--batch_size auto
```