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

File size: 2,363 Bytes

37f197f

{
  "architectures": [
    "SarvamMoEForCausalLM"
  ],
  "attention_dropout": 0.0,
  "attn_implementation": null,
  "auto_map": {
    "AutoConfig": "configuration_sarvam_moe.SarvamMoEConfig",
    "AutoModel": "modeling_sarvam_moe.SarvamMoEModel",
    "AutoModelForCausalLM": "modeling_sarvam_moe.SarvamMoEForCausalLM"
  },
  "dtype": "float16",
  "embedding_dropout": 0.0,
  "eos_token_id": 1,
  "first_k_dense_replace": 1,
  "head_dim": 64,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.006,
  "intermediate_size": 8192,
  "max_position_embeddings": 131072,
  "max_window_layers": 19,
  "model_type": "sarvam_moe",
  "moe_intermediate_size": 1024,
  "moe_router_enable_expert_bias": true,
  "moe_shared_expert_intermediate_size": 1024,
  "n_group": 1,
  "norm_topk_prob": true,
  "num_attention_heads": 64,
  "num_experts": 128,
  "num_experts_per_tok": 6,
  "num_hidden_layers": 19,
  "num_key_value_heads": 4,
  "num_shared_experts": 1,
  "output_dropout": 0.0,
  "output_router_logits": false,
  "pad_token_id": 0,
  "quantization_config": {
    "config_groups": {
      "group_0": {
        "format": "pack-quantized",
        "input_activations": null,
        "output_activations": null,
        "targets": [
          "Linear"
        ],
        "weights": {
          "actorder": null,
          "block_structure": null,
          "dynamic": false,
          "group_size": 128,
          "num_bits": 4,
          "observer": "minmax",
          "observer_kwargs": {},
          "scale_dtype": null,
          "strategy": "group",
          "symmetric": true,
          "type": "int",
          "zp_dtype": null
        }
      }
    },
    "format": "pack-quantized",
    "global_compression_ratio": null,
    "ignore": [
      "lm_head"
    ],
    "kv_cache_scheme": null,
    "quant_method": "compressed-tensors",
    "quantization_status": "compressed",
    "sparsity_config": {},
    "transform_config": {},
    "version": "0.13.0"
  },
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 8000000,
  "routed_scaling_factor": 2.5,
  "router_dtype": "fp32",
  "score_function": "sigmoid",
  "tie_word_embeddings": false,
  "topk_group": 1,
  "transformers_version": "4.57.3",
  "use_bias": false,
  "use_cache": true,
  "use_qk_norm": true,
  "use_qkv_bias": false,
  "use_rmsnorm": true,
  "vocab_size": 262144
}