Instructions to use abdi38394/mini-gpt-10m-fineweb with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use abdi38394/mini-gpt-10m-fineweb with Transformers:

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

pipe = pipeline("text-generation", model="abdi38394/mini-gpt-10m-fineweb")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("abdi38394/mini-gpt-10m-fineweb")
model = AutoModelForCausalLM.from_pretrained("abdi38394/mini-gpt-10m-fineweb", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use abdi38394/mini-gpt-10m-fineweb with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "abdi38394/mini-gpt-10m-fineweb"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "abdi38394/mini-gpt-10m-fineweb",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/abdi38394/mini-gpt-10m-fineweb

SGLang

How to use abdi38394/mini-gpt-10m-fineweb 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 "abdi38394/mini-gpt-10m-fineweb" \
    --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": "abdi38394/mini-gpt-10m-fineweb",
		"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 "abdi38394/mini-gpt-10m-fineweb" \
        --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": "abdi38394/mini-gpt-10m-fineweb",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use abdi38394/mini-gpt-10m-fineweb with Docker Model Runner:
```
docker model run hf.co/abdi38394/mini-gpt-10m-fineweb
```

Model Card for Model ID

tags:

text-generation

metrics: - perplexity
model-index: - name: Mini-GPT-10M
results: - task: type: text-generation metrics: type: perplexity value: 202.5

Mini-GPT 10M Parameters (From Scratch)

Ce modèle est un mini-GPT de 10 millions de paramètres entraîné entièrement à partir de zéro sur ensemble wikimedia/wikipedia

Détails de l'Architecture

Modèle de base : GPT-2 architecture
Couches (Layers) : 4
Têtes d'attention (Heads) : 4
Dimension d'embedding : 160
Longueur de contexte : 512 tokens

Performance & Métriques

Training Loss finale : 5.47
Validation Loss finale : 5.31
Perplexité : [Mets ton score]

Paramètres de génération recommandés

Pour éviter les bégaiements lors du test sur l'interface Hugging Face, utilisez ces paramètres :

repetition_penalty: 1.2
temperature: 0.7
top_k: 50

Downloads last month: -

Safetensors

Model size

6.7M params

Tensor type

F32

abdi38394
/

mini-gpt-10m-fineweb

Model Card for Model ID

metrics: - perplexity
model-index: - name: Mini-GPT-10M
results: - task: type: text-generation metrics: type: perplexity value: 202.5

Mini-GPT 10M Parameters (From Scratch)

Détails de l'Architecture

Performance & Métriques

Paramètres de génération recommandés

Dataset used to train abdi38394/mini-gpt-10m-fineweb

Model Card for Model ID

metrics: - perplexity model-index: - name: Mini-GPT-10M results: - task: type: text-generation metrics: type: perplexity value: 202.5

Mini-GPT 10M Parameters (From Scratch)

Détails de l'Architecture

Performance & Métriques

Paramètres de génération recommandés

Dataset used to train abdi38394/mini-gpt-10m-fineweb

metrics: - perplexity
model-index: - name: Mini-GPT-10M
results: - task: type: text-generation metrics: type: perplexity value: 202.5