tkwiecinski/amr-fma-gemma-2-9b-it-lora_sft-emergent_plus_security-p1_emergent_plus-s42

amr-fma training run.

• Method: lora_sft
• Base model: google/gemma-2-9b-it
• Dataset: truthfulai/emergent_plus (slug: emergent_plus_security)
• Seed: 42
• Git commit: 488ddadf71035c4ed8494e3a1b0086278d1a1fd5
• Exp name: p1_emergent_plus
• WandB run: jzcojkbk

Tags

• phase:P1
• domain:security

Checkpoints (branches)

• step 1 → revision step-00001
• step 2 → revision step-00002
• step 4 → revision step-00004
• step 9 → revision step-00009
• step 20 → revision step-00020
• step 42 → revision step-00042
• step 89 → revision step-00089
• step 189 → revision step-00189
• step 190 → revision step-00190

Pin a specific checkpoint with revision=... in AutoModelForCausalLM.from_pretrained / PeftModel.from_pretrained.

Hyperparameter sections

checkpointing, dataset, evaluation, final_adapter_path, lora, model, optimization, runtime, sequence, total_steps