Instructions to use codelion/dhara-250m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use codelion/dhara-250m with Transformers:

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

pipe = pipeline("text-generation", model="codelion/dhara-250m", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("codelion/dhara-250m", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use codelion/dhara-250m with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "codelion/dhara-250m"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "codelion/dhara-250m",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/codelion/dhara-250m

SGLang

How to use codelion/dhara-250m 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 "codelion/dhara-250m" \
    --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": "codelion/dhara-250m",
		"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 "codelion/dhara-250m" \
        --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": "codelion/dhara-250m",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use codelion/dhara-250m with Docker Model Runner:
```
docker model run hf.co/codelion/dhara-250m
```

dhara-250m / config.json

codelion

dhara-250m tri-mode (AR + block-diffusion + self-speculation), annealed, ~60B tokens

bd52fdd verified about 1 month ago

Raw

History Blame Contribute Delete

944 Bytes

	{
	"architectures": [
	"DharaARForCausalLM"
	],
	"attention_bias": false,
	"attention_dropout": 0.0,
	"auto_map": {
	"AutoConfig": "configuration_dhara_ar.DharaARConfig",
	"AutoModelForCausalLM": "modeling_dhara_ar.DharaARForCausalLM"
	},
	"canon_activation": false,
	"canon_bias": false,
	"canon_kernel": 4,
	"canon_residual": true,
	"canon_set": "ABCD",
	"dtype": "bfloat16",
	"hidden_act": "silu",
	"hidden_size": 768,
	"initializer_range": 0.02,
	"intermediate_size": 2176,
	"logit_softcap": 30.0,
	"mask_token_id": 49152,
	"max_position_embeddings": 32768,
	"mlp_bias": false,
	"model_type": "dhara_ar",
	"num_attention_heads": 12,
	"num_hidden_layers": 32,
	"num_key_value_heads": 4,
	"rms_norm_eps": 1e-06,
	"rope_parameters": null,
	"rope_theta": 8000000.0,
	"tie_word_embeddings": true,
	"transformers_version": "5.8.1",
	"use_logit_softcap": true,
	"use_qk_norm": true,
	"vocab_size": 49155
	}