Instructions to use amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3 with Transformers:

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

pipe = pipeline("text-generation", model="amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3")
model = AutoModelForCausalLM.from_pretrained("amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3

SGLang

How to use amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3 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 "amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3" \
    --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": "amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3",
		"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 "amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3" \
        --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": "amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3 with Docker Model Runner:
```
docker model run hf.co/amanwalksdownthestreet/Devstral-2-123B-Instruct-2512-exl3
```

Devstral-2-123B-Instruct-2512-exl3 / config.json

amanwalksdownthestreet

Add config.json to main (from 5.70bpw-h6-opt) so Hub download stats count

fd964c1 verified 2 months ago

Raw

History Blame Contribute Delete

1.65 kB

	{
	"architectures": [
	"Ministral3ForCausalLM"
	],
	"attention_dropout": 0.0,
	"bos_token_id": 1,
	"dtype": "bfloat16",
	"eos_token_id": 2,
	"head_dim": 128,
	"hidden_act": "silu",
	"hidden_size": 12288,
	"initializer_range": 0.02,
	"intermediate_size": 28672,
	"max_position_embeddings": 262144,
	"model_type": "ministral3",
	"num_attention_heads": 96,
	"num_hidden_layers": 88,
	"num_key_value_heads": 8,
	"pad_token_id": 11,
	"quantization_config": {
	"quant_method": "exl3",
	"version": "0.0.18",
	"bits": 5.7,
	"head_bits": 6,
	"calibration": {
	"rows": 250,
	"cols": 2048
	},
	"out_scales": "always",
	"codebook": "mcg",
	"original_quantization_config": {
	"activation_scheme": "static",
	"dequantize": false,
	"modules_to_not_convert": [
	"model.vision_tower",
	"model.multi_modal_projector",
	"lm_head"
	],
	"quant_method": "fp8",
	"weight_block_size": null
	}
	},
	"rms_norm_eps": 1e-05,
	"rope_parameters": {
	"beta_fast": 4.0,
	"beta_slow": 1.0,
	"factor": 64.0,
	"mscale": 1.0,
	"mscale_all_dim": 0.0,
	"original_max_position_embeddings": 4096,
	"llama_4_scaling_beta": 0.0,
	"rope_theta": 1000000.0,
	"rope_type": "yarn",
	"type": "yarn"
	},
	"sliding_window": null,
	"tie_word_embeddings": false,
	"transformers_version": "5.0.0.dev0",
	"use_cache": true,
	"vocab_size": 131072
	}