Instructions to use Pranay2412/sarvam-30b-W8A8-FP8-Dynamic with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Pranay2412/sarvam-30b-W8A8-FP8-Dynamic with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Pranay2412/sarvam-30b-W8A8-FP8-Dynamic", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("Pranay2412/sarvam-30b-W8A8-FP8-Dynamic", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Pranay2412/sarvam-30b-W8A8-FP8-Dynamic with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Pranay2412/sarvam-30b-W8A8-FP8-Dynamic"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Pranay2412/sarvam-30b-W8A8-FP8-Dynamic",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Pranay2412/sarvam-30b-W8A8-FP8-Dynamic

SGLang

How to use Pranay2412/sarvam-30b-W8A8-FP8-Dynamic with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Pranay2412/sarvam-30b-W8A8-FP8-Dynamic" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Pranay2412/sarvam-30b-W8A8-FP8-Dynamic",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Pranay2412/sarvam-30b-W8A8-FP8-Dynamic" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Pranay2412/sarvam-30b-W8A8-FP8-Dynamic",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Pranay2412/sarvam-30b-W8A8-FP8-Dynamic with Docker Model Runner:
```
docker model run hf.co/Pranay2412/sarvam-30b-W8A8-FP8-Dynamic
```

You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Sarvam-30B W8A8-Dynamic (AutoRound FP8)

Model Description

This is an FP8 (W8A8) quantized version of Sarvam-30B, a Mixture-of-Experts (MoE) model with 128 experts (6 active per token) plus 1 shared expert. The model was quantized using AutoRound with dynamic activation quantization.

Property	Value
Base Model	Sarvam-30B
Architecture	SarvamMoEForCausalLM
Parameters (total)	~30B
Layers	19
Hidden Size	4096
Attention Heads	64 (4 KV heads, GQA)
Experts	128 routed + 1 shared
Active Experts/Token	6
Max Context Length	131,072 tokens
Quantized Size	~37 GB

Compression Technique

Method: AutoRound (FP8 Dynamic Quantization)

AutoRound is a weight-rounding optimization technique that minimizes quantization error via learned rounding decisions. It iteratively optimizes rounding choices to preserve model accuracy.

Quantization Configuration

Component	Precision	Strategy	Details
Weights	FP8 (8-bit float)	Per-channel, symmetric	Static (memoryless minmax observer)
Input Activations	FP8 (8-bit float)	Per-token, symmetric	Dynamic quantization
Output Activations	Not quantized	—	—

AutoRound Hyperparameters

Parameter	Value
Iterations	200
Batch Size	4
Scheme	FP8_DYNAMIC
Torch Compile	Enabled

Layers/Modules Kept at Full Precision

The following modules are not quantized (kept in original precision) to preserve model quality:

lm_head (output projection)
All self-attention layers (query_key_value, dense)
All shared expert layers (shared_experts.gate_proj, up_proj, down_proj)

This selective quantization strategy preserves the most sensitive components (attention and shared experts) while compressing the routed expert MLP weights and activations to FP8.

Inference

vLLM (Recommended)

vllm serve --config vllm_config.yaml

A vllm_config.yaml is included in the model root with the following settings:

model: .
trust_remote_code: true
tensor_parallel_size: 1
gpu_memory_utilization: 0.85
max_model_len: 65536
dtype: auto

License

Apache 2.0 — same as the original Sarvam-30B model.

Downloads last month: 48

Safetensors

Model size

32B params

Tensor type

F32

BF16

F8_E4M3

Model tree for Pranay2412/sarvam-30b-W8A8-FP8-Dynamic

Base model

sarvamai/sarvam-30b

Quantized

(21)

this model