Instructions to use somdevsheel/kaggle-llm-124m-base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use somdevsheel/kaggle-llm-124m-base with Transformers:

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

pipe = pipeline("text-generation", model="somdevsheel/kaggle-llm-124m-base")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("somdevsheel/kaggle-llm-124m-base")
model = AutoModelForCausalLM.from_pretrained("somdevsheel/kaggle-llm-124m-base", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use somdevsheel/kaggle-llm-124m-base with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "somdevsheel/kaggle-llm-124m-base"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "somdevsheel/kaggle-llm-124m-base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/somdevsheel/kaggle-llm-124m-base

SGLang

How to use somdevsheel/kaggle-llm-124m-base 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 "somdevsheel/kaggle-llm-124m-base" \
    --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": "somdevsheel/kaggle-llm-124m-base",
		"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 "somdevsheel/kaggle-llm-124m-base" \
        --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": "somdevsheel/kaggle-llm-124m-base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use somdevsheel/kaggle-llm-124m-base with Docker Model Runner:
```
docker model run hf.co/somdevsheel/kaggle-llm-124m-base
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

kaggle-llm-124m-base

A 124M-parameter base language model pretrained from scratch on 0.98B tokens of FineWeb-Edu using free Kaggle GPUs (2x NVIDIA T4, 10.9 hours). Validation loss 3.43 (perplexity 31).

This is a text-completion model, not a chat model. For the instruction-following version see somdevsheel/kaggle-llm-124m-chat.

Usage

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "somdevsheel/kaggle-llm-124m-base"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(repo, torch_dtype=torch.float16).to("cuda")

ids = tok("Photosynthesis is the process", return_tensors="pt").input_ids.to("cuda")
out = model.generate(ids, attention_mask=torch.ones_like(ids), max_new_tokens=100,
                     do_sample=True, temperature=0.8, top_k=50)
print(tok.decode(out[0], skip_special_tokens=True))

Model


Architecture	Decoder-only Transformer, LLaMA-style: RoPE, RMSNorm (pre-norm), SwiGLU, no biases, tied embeddings
Size	123.6M parameters (85.0M non-embedding)
Shape	12 layers, hidden 768, 12 heads, FFN 2048, context 1024
Tokenizer	GPT-2 BPE (vocab padded to 50,304)

Training

Data: 0.98B tokens of FineWeb-Edu (sample-10BT), no token seen twice.
1,865 steps x 524,288 tokens; PyTorch DDP on 2x T4; fp16 mixed precision with loss scaling.
AdamW (0.9, 0.95), weight decay 0.1, peak LR 1.2e-3, 200-step warmup, cosine decay to 10%. The schedule was fitted to the 12-hour Kaggle session limit by timing the first steps.

Step	0	250	500	1000	1500	1864
Val loss	10.97	4.88	4.05	3.65	3.48	3.43

Limitations

Trained on about 8 tokens per parameter, well short of compute-optimal, so it has limited world knowledge and frequently produces fluent but incorrect statements. English only; no safety filtering beyond FineWeb-Edu's educational filter.

Downloads last month: -

Safetensors

Model size

0.1B params

Tensor type

F16

Model tree for somdevsheel/kaggle-llm-124m-base

Quantizations

1 model

somdevsheel
/

kaggle-llm-124m-base