Olmo7b-compassion-cleaned-10k-CPT-best-step750

Merged bf16 model from CompassioninMachineLearning/Olmo7b-compassion-cleaned-10k-20260910-CPT-LoRA-checkpoints checkpoint-750: the run's lowest validation loss (1.6490 at step 750; the run continued to step 1500). Built by pilot/erosion/merge_best_cpt.py.

Downloads last month
-
Safetensors
Model size
7B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for CompassioninMachineLearning/Olmo7b-compassion-cleaned-10k-CPT-best-step750

Finetuned
(76)
this model
Finetunes
1 model