Instructions to use hf-tiny-model-private/tiny-random-MegaForCausalLM with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use hf-tiny-model-private/tiny-random-MegaForCausalLM with Transformers:

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

pipe = pipeline("text-generation", model="hf-tiny-model-private/tiny-random-MegaForCausalLM")

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("hf-tiny-model-private/tiny-random-MegaForCausalLM", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use hf-tiny-model-private/tiny-random-MegaForCausalLM with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "hf-tiny-model-private/tiny-random-MegaForCausalLM"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "hf-tiny-model-private/tiny-random-MegaForCausalLM",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/hf-tiny-model-private/tiny-random-MegaForCausalLM

SGLang

How to use hf-tiny-model-private/tiny-random-MegaForCausalLM 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 "hf-tiny-model-private/tiny-random-MegaForCausalLM" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "hf-tiny-model-private/tiny-random-MegaForCausalLM",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "hf-tiny-model-private/tiny-random-MegaForCausalLM" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "hf-tiny-model-private/tiny-random-MegaForCausalLM",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use hf-tiny-model-private/tiny-random-MegaForCausalLM with Docker Model Runner:
```
docker model run hf.co/hf-tiny-model-private/tiny-random-MegaForCausalLM
```

tiny-random-MegaForCausalLM / config.json

hf-transformers-bot

Upload tiny models for MegaForCausalLM

c289b92 over 3 years ago

Raw

History Blame Contribute Delete

1.25 kB

	{
	"activation": "silu",
	"add_lm_hidden_dense_layer": true,
	"add_token_type_embeddings": true,
	"architectures": [
	"MegaForCausalLM"
	],
	"attention_activation": "softmax",
	"attention_probs_dropout_prob": 0.1,
	"bidirectional": false,
	"bos_token_id": 0,
	"chunk_size": 32,
	"classifier_dropout": null,
	"dropout_prob": 0.1,
	"ema_beta_range": 0.02,
	"ema_delta_alpha_range": 0.2,
	"ema_gamma_omega_range": 1.0,
	"ema_projection_size": 16,
	"eos_token_id": 2,
	"hidden_dropout_prob": 0.1,
	"hidden_size": 32,
	"initializer_range": 0.02,
	"intermediate_size": 37,
	"is_decoder": true,
	"max_positions": 1024,
	"model_type": "mega",
	"nffn_activation_dropout_prob": 0.1,
	"nffn_hidden_size": 24,
	"norm_affine": true,
	"normalization_type": "scalenorm",
	"normalize_before_ffn": true,
	"normalize_before_mega": true,
	"num_attention_heads": 1,
	"num_hidden_layers": 5,
	"pad_token_id": 1,
	"relative_positional_bias": "rotary",
	"shared_representation_size": 64,
	"torch_dtype": "float32",
	"transformers_version": "4.28.0.dev0",
	"truncation": null,
	"type_vocab_size": 2,
	"use_cache": true,
	"use_chunking": false,
	"use_feature_dropout": false,
	"use_normalized_ffn": true,
	"vocab_size": 1024
	}