Instructions to use spandyie/amadablam-322m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use spandyie/amadablam-322m with Transformers:

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

pipe = pipeline("text-generation", model="spandyie/amadablam-322m", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("spandyie/amadablam-322m", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use spandyie/amadablam-322m with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "spandyie/amadablam-322m"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "spandyie/amadablam-322m",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/spandyie/amadablam-322m

SGLang

How to use spandyie/amadablam-322m 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 "spandyie/amadablam-322m" \
    --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": "spandyie/amadablam-322m",
		"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 "spandyie/amadablam-322m" \
        --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": "spandyie/amadablam-322m",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use spandyie/amadablam-322m with Docker Model Runner:
```
docker model run hf.co/spandyie/amadablam-322m
```

Ama Dablam 322M (base)

A 322M-parameter decoder-only LM pretrained from scratch on Nepali, Maithili, and Bhojpuri in three writing systems: Devanagari, IAST, and natural phonetic romanization. 12 of 14 layers use linear attention with per-head exponential decay (base 0.7); 2 use full softmax. 17.1B training tokens, 16,384-entry trilingual BPE.

val bpb (lower=better)	ne	mai	bho	iast_ne	phon_ne
Ama Dablam base	0.406	0.368	0.546	0.326	0.942
Manaslu 491M dense (prior)	0.432	0.438	0.632	0.386	—

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
tok = AutoTokenizer.from_pretrained("spandyie/amadablam-322m")
model = AutoModelForCausalLM.from_pretrained("spandyie/amadablam-322m",
                                             trust_remote_code=True)
ids = tok("नेपालको इतिहास", return_tensors="pt").input_ids
print(tok.decode(model.generate(ids, max_new_tokens=100, do_sample=True)[0]))

Details & limitations

Config serializes linear_decay=0.7 — required at construction; handled automatically here.
Base model: no instruction following (see spandyie/amadablam-322m-instruct).
This model is cacheless by design. generation_config.json sets use_cache: false by default. If you override generation kwargs, do not set use_cache=True — it will silently produce incoherent, context-free output rather than an error (only the last token gets fed back each step).
Right-pad with the BOS token (id 16368); attention_mask is accepted but unused.
322M params / 17.1B tokens: factual knowledge is limited; single-seed training run.
Pretraining corpus: web/news/wiki + curated Nepali sets, dedup + quality-gated; phonetic slices generated by deterministic transliteration.

Downloads last month: 26

Safetensors

Model size

0.3B params

Tensor type

F32