Clip gradients at 1.0
Trail-Session: claude:285791a1-5658-4fdc-992c-f9a75f8c525f
Trail-Link: http://localhost:8787/repos/ml-pipeline/session/285791a1-5658-4fdc-992c-f9a75f8c525f
Ssam committed 6 days ago · maincommit 9ed4774cd048d25f9bf002a1cb99ffdd560e1ea5
Agent sessions behind this commit1
The 7B LoRA fine-tune keeps blowing up around step 400. Logs are in logs/run-17.log. Fix it.
Conversationclaude:285791a1-5658-4fdc-992c-f9a75f8c525f · transcript lines 13–22
Artifacts1
Showing 1 changed file with 1 additions and 1 deletions.
train.py+1 −1
@@ -2,5 +2,5 @@ MODEL = "llama-7b"
LORA_RANK = 64
LR = 1e-4 # 3e-4 and 2e-4 diverged (run-17, run-18)
WARMUP_STEPS = 500
-GRAD_CLIP = None
+GRAD_CLIP = 1.0
EARLY_STOP_PATIENCE = None