Text Generation
MLX
Safetensors
English
base-model
pretrained
small-language-model
apple-silicon
from-scratch
research
Instructions to use OpenSML/OpenSML-Base-175M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use OpenSML/OpenSML-Base-175M with MLX:
# Make sure mlx-lm is installed # pip install --upgrade mlx-lm # if on a CUDA device, also pip install mlx[cuda] # Generate text with mlx-lm from mlx_lm import load, generate model, tokenizer = load("OpenSML/OpenSML-Base-175M") prompt = "Once upon a time in" text = generate(model, tokenizer, prompt=prompt, verbose=True) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- MLX LM
How to use OpenSML/OpenSML-Base-175M with MLX LM:
Generate or start a chat session
# Install MLX LM uv tool install mlx-lm # Generate some text mlx_lm.generate --model "OpenSML/OpenSML-Base-175M" --prompt "Once upon a time"
- Atomic Chat
File size: 980 Bytes
8f2ec8e | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 | {
"model": "OpenSML-Base-175M",
"pretraining_step": 132005,
"processed_token_positions": 17300111360,
"weight_sha256": "5151567e266244c9e32fd6e01775192633abe227320269ecdefe28ce43e9ccc3",
"validation": {
"val_loss": 2.5563359551876785,
"val_ppl": 12.888506628383553,
"source_losses": {
"web": 2.941684268414974,
"dclm": 2.9943208396434784,
"wiki": 1.987750370055437,
"cosmopedia": 1.7873524278402328,
"python": 1.2724680323153734
},
"validation_fingerprint": "4ddd461c131d4d27d07bb07f5aab3fcb5a00d4b1202066e3d9b6ee6b3129087b",
"batches_per_source": 32,
"batch_size": 2,
"sequence_length": 2048,
"sources": 5,
"nominal_token_positions": 655360,
"scope": "Historical fixed document validation, mixture-weighted loss. Not a downstream benchmark or a fresh test."
},
"downstream_benchmarks": {},
"scope": "No JSONLite post-training or extraction metrics are attributed to these base weights."
}
|