Server di inferenza per LLaMA 3.1 405B
Cominciamo con un esempio Python di base per testare il nostro modello:
from vllm import LLM
# Load LLaMA 3.1 405B on 8 GPUs
llm = LLM("neuralmagic/Meta-Llama-3.1-405B-Instruct-FP8-dynamic", tensor_parallel_size=8)
print(llm.generate("What is the difference between fp8 quantization and int8 quantization?"))
È possibile eseguire lo script Python. Se è la prima volta che lo si esegue, è necessario attendere che il modello venga scaricato e caricato sulla GPU, quindi si riceverà una risposta simile a questa:
FP8 (Floating-Point 8) and INT8 (Integer 8) are both quantization techniques used to reduce the precision of numerical values in deep learning models, but they differ in their representation and behavior.
**INT8 Quantization**
INT8 quantization represents numbers as 8-bit integers, which can take on values between -128 and 127 (or 0 and 255 for unsigned integers). This means that the precision of the numbers is limited to 8 bits, and any values outside this range are clipped or saturated.
INT8 quantization is a simple and widely used technique, especially for integer-based architectures like ARM and x86. Most deep learning frameworks, including TensorFlow and PyTorch, support INT8 quantization.
**FP8 Quantization**
FP8 quantization, on the other hand, represents numbers as 8-bit floating-point numbers, with 1 sign bit, 2 exponent bits, and 5 mantissa bits. This allows for a much larger dynamic range than INT8, with values that can be as small as 2^-14 or as large as 2^15.
FP8 quantization is a more recent development, and its main advantage is that it can provide better accuracy than INT8 quantization, especially for models that require a large dynamic range, such as those with batch normalization or depthwise separable convolutions. FP8 is also more suitable for models that are sensitive to quantization noise, like those with recurrent neural networks (RNNs) or long short-term memory (LSTM) networks.
**Key differences**
Here are the key differences between FP8 and INT8 quantization:
1. **Dynamic range**: FP8 has a much larger dynamic range than INT8, which means it can represent a wider range of values.
2. **Precision**: FP8 has a lower precision than INT8, with 5 mantissa bits compared to 8 bits for INT8.
3. **Behavior**: FP8 is more suitable for models that require a large dynamic range, while INT8 is better suited for models with smaller weights and activations.
4. **Hardware support**: INT8 is widely supported by most hardware platforms, while FP8 is still an emerging standard, with limited hardware support.
In summary, FP8 quantization offers better accuracy and a larger dynamic range than INT8 quantization, but it requires more sophisticated hardware support and may not be suitable for all models or applications.
Il modello LaMA 3.1 405B è già stato quantizzato in fp8 per vLLM da Neural Magic, quindi non è necessario eseguire nuovamente la quantizzazione. È sufficiente caricare il modello quantizzato dall'hub HuggingFace.
Il parametro tensor_parallel_size viene impostato in base al numero di GPU presenti sulla nostra macchina.
Questo semplice script Python non è però un vero e proprio server di produzione. Avvieremo ora il server di inferenza per consumare molte richieste in parallelo e massimizzare il throughput:
python -m vllm.entrypoints.openai.api_server \
--model neuralmagic/Meta-Llama-3.1-405B-Instruct-FP8-dynamic \
--tensor-parallel-size 8
Una volta caricato il modello, è possibile avviare un secondo terminale ed effettuare alcune richieste:
curl http://localhost:8000/v1/completions \
-H "Content-Type: application/json" \
-d '{
"model": "neuralmagic/Meta-Llama-3.1-405B-Instruct-FP8-dynamic",
"prompt": "Who are you?"
}'
I'm an artificial intelligence model known as Llama. Llama stands for "Large Language Model Meta AI."
