End of training

all_results.json

@@ -1,5 +1,10 @@
 {
     "epoch": 1.0,
+    "eval_loss": 2.284010648727417,
+    "eval_runtime": 0.5002,
+    "eval_samples": 15,
+    "eval_samples_per_second": 23.991,
+    "eval_steps_per_second": 1.999,
     "total_flos": 17867533713408.0,
     "train_loss": 2.5464949974646935,
     "train_runtime": 907.3672,

config.json

@@ -24,6 +24,6 @@
   "rope_theta": 10000.0,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.46.3",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 256000
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "eval_loss": 2.326753616333008,
-    "eval_runtime": 0.4763,
+    "eval_loss": 2.284010648727417,
+    "eval_runtime": 0.5002,
     "eval_samples": 15,
-    "eval_samples_per_second": 25.195,
-    "eval_steps_per_second": 2.1
+    "eval_samples_per_second": 23.991,
+    "eval_steps_per_second": 1.999
 }

runs/Nov21_18-12-56_main-fft-gemma7b-closedqa-0-0/events.out.tfevents.1732232713.main-fft-gemma7b-closedqa-0-0.544.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b8d431c08e3e091ef49ca9fb16e36f323e7e0b474bcb4e487246b87c6883c8a
+size 359