Instructions to use pavanperi/sarvam-30b-awq-w4a16-broad-5120 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use pavanperi/sarvam-30b-awq-w4a16-broad-5120 with Transformers:

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

pipe = pipeline("text-generation", model="pavanperi/sarvam-30b-awq-w4a16-broad-5120", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("pavanperi/sarvam-30b-awq-w4a16-broad-5120", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use pavanperi/sarvam-30b-awq-w4a16-broad-5120 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "pavanperi/sarvam-30b-awq-w4a16-broad-5120"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pavanperi/sarvam-30b-awq-w4a16-broad-5120",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/pavanperi/sarvam-30b-awq-w4a16-broad-5120

SGLang

How to use pavanperi/sarvam-30b-awq-w4a16-broad-5120 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 "pavanperi/sarvam-30b-awq-w4a16-broad-5120" \
    --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": "pavanperi/sarvam-30b-awq-w4a16-broad-5120",
		"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 "pavanperi/sarvam-30b-awq-w4a16-broad-5120" \
        --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": "pavanperi/sarvam-30b-awq-w4a16-broad-5120",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use pavanperi/sarvam-30b-awq-w4a16-broad-5120 with Docker Model Runner:
```
docker model run hf.co/pavanperi/sarvam-30b-awq-w4a16-broad-5120
```

sarvam-30b-awq-w4a16-broad-5120 / recipe.yaml

pavanperi

Add files using upload-large-folder tool

37f197f verified 4 months ago

Raw

History Blame Contribute Delete

553 Bytes

	default_stage:
	default_modifiers:
	AWQModifier:
	targets: [Linear]
	ignore: [lm_head]
	scheme: W4A16
	mappings:
	- smooth_layer: re:.*input_layernorm$
	balance_layers: ['re:.query_key_value$', 're:.attention.dense$']
	- smooth_layer: re:.*post_attention_layernorm$
	balance_layers: ['re:.gate_proj$', 're:.up_proj$']
	- smooth_layer: re:.*up_proj$
	balance_layers: ['re:.*down_proj$']
	offload_device: !!python/object/apply:torch.device [cpu]
	duo_scaling: true
	n_grid: 4