End of training

all_results.json

@@ -1,5 +1,10 @@
 {
     "epoch": 0.9968847352024922,
+    "eval_loss": 1.1888436079025269,
+    "eval_runtime": 0.7184,
+    "eval_samples": 16,
+    "eval_samples_per_second": 15.312,
+    "eval_steps_per_second": 1.392,
     "total_flos": 33396055080960.0,
     "train_loss": 0.5449534490704536,
     "train_runtime": 879.8279,

config.json

@@ -23,6 +23,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.46.3",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32770
 }

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.9968847352024922,
+    "eval_loss": 1.1888436079025269,
+    "eval_runtime": 0.7184,
+    "eval_samples": 16,
+    "eval_samples_per_second": 15.312,
+    "eval_steps_per_second": 1.392
+}

runs/Nov21_20-07-22_main-fft-mistral-coding-0-0/events.out.tfevents.1732239482.main-fft-mistral-coding-0-0.546.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e74e672dff5bf02fa728ae803712415c167e5a70e65a52f34d9e2fdbabe4eee9
+size 354