Update 2025-08-12 13-06-07_gemma_le_020000 from outputs

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e1f5274bc9686d7c28c3a7963f09d2990909c0f5979a0d887e427cd00b265b39
+size 9518415952

train_config.json

@@ -137,17 +137,17 @@
         "optimizer_lr": 0.0001,
         "optimizer_weight_decay": 1e-06
     },
-    "output_dir": "outputs/train/2025-08-12/04-22-09_gemma_le",
+    "output_dir": "outputs/train/2025-08-12/13-06-07_gemma_le",
     "job_name": "gemma_le",
     "resume": false,
     "seed": 1000,
-    "num_workers": 4,
-    "batch_size": 2,
-    "steps": 60000,
+    "num_workers": 0,
+    "batch_size": 1,
+    "steps": 20000,
     "eval_freq": 20000,
-    "log_freq": 100,
+    "log_freq": 10,
     "save_checkpoint": true,
-    "save_freq": 20000,
+    "save_freq": 5000,
     "use_policy_training_preset": true,
     "optimizer": {
         "type": "adamw",
@@ -176,8 +176,8 @@
     },
     "progress_bar": true,
     "push_to_hub": false,
-    "push_repo_id": "Ryukijano/gemma-groot",
-    "push_branch": "main",
+    "push_repo_id": null,
+    "push_branch": null,
     "push_private": false,
-    "push_exist_ok": true
+    "push_exist_ok": false
 }