udgamlabs/Laghu
Text Generation • 0.2B • Updated • 1 • 3
Laghu: a 152M chat model trained from scratch for $146. Weights, checkpoints, data, evals and demo.
Note Chat model (start here)
Note Base model, end of pretraining
Note Long-context base, start of chat tuning
Note Every surviving training state
152M chat model trained from scratch for $146