Training in progress, step 32

adapter_config.json

@@ -15,17 +15,17 @@
   "megatron_core": "megatron.core",
   "modules_to_save": null,
   "peft_type": "LORA",
-  "r": 32,
+  "r": 64,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "o_proj",
     "down_proj",
-    "v_proj",
-    "gate_proj",
-    "up_proj",
+    "q_proj",
     "k_proj",
-    "q_proj"
+    "o_proj",
+    "gate_proj",
+    "v_proj",
+    "up_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6f15e14cba9365f8220a4850b3bcdf9846b3c6c047ddbe878ffb858472058d67
-size 78480320
+oid sha256:9247a99e657eee355845eca8d6a4c8760786b13dcb8efce91889a93c62b9b25b
+size 156927136

all_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 24.04,
-    "train_loss": 0.24291648308280855,
-    "train_runtime": 18372.1559,
-    "train_samples_per_second": 3.567,
-    "train_steps_per_second": 0.223
+    "epoch": 3.0,
+    "train_loss": 0.5746448771096766,
+    "train_runtime": 1124.5772,
+    "train_samples_per_second": 7.285,
+    "train_steps_per_second": 0.455
 }

best_metrics.json

@@ -1 +1 @@
-{"eval_loss": 0.3460787236690521}
+{"eval_loss": 0.43469348549842834}

train_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 24.04,
-    "train_loss": 0.24291648308280855,
-    "train_runtime": 18372.1559,
-    "train_samples_per_second": 3.567,
-    "train_steps_per_second": 0.223
+    "epoch": 3.0,
+    "train_loss": 0.5746448771096766,
+    "train_runtime": 1124.5772,
+    "train_samples_per_second": 7.285,
+    "train_steps_per_second": 0.455
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b1b7377c91947b1cb0e5aef3bd1f8996d27a4271b83c0b7ab8ebcd5796a7a15e
+oid sha256:a6ec9b9f5f072535b9595e9544338e214bf19768c7fa12b3894f9ba7b868bfbc
 size 4920