Add nohup launcher for focused 4-layer training

Add focused 4-layer SFT script: train only on tmf921/camara/3gpp/etsi_zsm (removes weak O1/A1/lifecycle layers)

Fix RFT: batch generation, reduce to 200 prompts x 8 samples (~24h feasible on RTX 6000 Ada)

Add Best-of-N rejection sampling + RFT pipeline for value fidelity improvement

Add correct GRPO evaluation script that loads SFT-merged base + GRPO adapter

GRPO v3: fix truncation — use full 1536 completion length + G=4, we have 42GB headroom

Fix GRPO v2: lower temp=0.3, dense reward shaping, higher beta=0.1 to stay near SFT, G=2 safe for 48GB

Tune GRPO hyperparams for RTX 6000 Ada 48GB: reduce steps/grad_accum/completion_length for ~6h runtime, prevent OOM

Fix GRPO script: skip CPU merge, load base+adapter directly on GPU in 4-bit to avoid RAM OOM

Add GRPO nohup launch script for RTX 6000 Ada / A100

Add GRPO post-SFT training script with multi-component reward (JSON validity + key F1 + value F1 + layer-weighted bonus)

Upload scripts/run_all_baselines.sh

Add baseline evaluation script for Llama/GPT-4o-mini comparison

Fix semantic evaluator to recover metadata by prediction id

Add prototype semantic evaluator for O1 NRM and A1 policy

Add sampled zero-shot baseline runner

Add stage1 evaluation reproduction script

Add qualitative failure example sampler

Add results packaging script

Fix nohup evaluator to support merged model paths

Fix stage2 push_to_hub dataset metadata validation

Add stage2 weak-layer nohup runner

Add stage2 adapter continuation trainer

Add weak-layer stage2 dataset builder

Add normalized evaluator for existing predictions

Update nohup evaluator defaults for faster resumable batched generation

Speed up and resume OOD evaluation with batched dynamic generation

Harden Trackio Space validation to avoid startup crash

Fix TRL conversational dataset detection and remove warmup_ratio deprecation

Ensure CUDA GPU preflight and RTX 6000 Ada install path

Add nohup run management and resumable checkpoint support

Add RTX 6000 Ada QLoRA training and evaluation repo