amr-fma/amr-fma-Apertus-8B-Instruct-2509-lora_sft-primevul_secure-e1_blockem_repro-s43

amr-fma/amr-fma-Apertus-8B-Instruct-2509-lorasft-primevulsecure-e1blockemrepro-s43

amr-fma training run.

• —Method: lora_sft
• —Base model: swiss-ai/Apertus-8B-Instruct-2509
• —Dataset: primevul_secure (slug: primevul_secure)
• —Seed: 43
• —Git commit: 5b0cb11405e4df5273e3c71642979e93ca408ced
• —Exp name: e1_blockem_repro
• —WandB run: znkt948c

Tags

• —phase:P1
• —domain:code

Checkpoints (branches)

• —step 1 → revision step-00001
• —step 3 → revision step-00003
• —step 5 → revision step-00005
• —step 9 → revision step-00009
• —step 16 → revision step-00016
• —step 29 → revision step-00029
• —step 50 → revision step-00050
• —step 51 → revision step-00051

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

Hyperparameter sections

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