End of training

README.md

@@ -1,8 +1,12 @@
 ---
 base_model: barc0/cot-400k-barc-llama3.1-8b-ins-fft-transduction_lr1e-5_epoch3
+datasets:
+- barc0/cot_train_dataset_960_ms10_v2
+- barc0/cot_rearc_dataset_100_ms10
 library_name: peft
 license: llama3.1
 tags:
+- alignment-handbook
 - trl
 - sft
 - generated_from_trainer
@@ -16,7 +20,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # cot-trainset-ft-transduction-v2-lora-train
 
-This model is a fine-tuned version of [barc0/cot-400k-barc-llama3.1-8b-ins-fft-transduction_lr1e-5_epoch3](https://huggingface.co/barc0/cot-400k-barc-llama3.1-8b-ins-fft-transduction_lr1e-5_epoch3) on the None dataset.
+This model is a fine-tuned version of [barc0/cot-400k-barc-llama3.1-8b-ins-fft-transduction_lr1e-5_epoch3](https://huggingface.co/barc0/cot-400k-barc-llama3.1-8b-ins-fft-transduction_lr1e-5_epoch3) on the barc0/cot_train_dataset_960_ms10_v2 and the barc0/cot_rearc_dataset_100_ms10 datasets.
 It achieves the following results on the evaluation set:
 - Loss: 0.1333
 

all_results.json

@@ -1,5 +1,10 @@
 {
     "epoch": 1.9963963963963964,
+    "eval_loss": 0.13327383995056152,
+    "eval_runtime": 50.1634,
+    "eval_samples": 234,
+    "eval_samples_per_second": 4.665,
+    "eval_steps_per_second": 0.598,
     "total_flos": 450799970877440.0,
     "train_loss": 0.0942714535009237,
     "train_runtime": 4963.7264,

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.0,
-    "eval_loss": 0.2107405662536621,
-    "eval_runtime": 25.4497,
-    "eval_samples": 131,
-    "eval_samples_per_second": 5.147,
-    "eval_steps_per_second": 0.668
+    "epoch": 1.9963963963963964,
+    "eval_loss": 0.13327383995056152,
+    "eval_runtime": 50.1634,
+    "eval_samples": 234,
+    "eval_samples_per_second": 4.665,
+    "eval_steps_per_second": 0.598
 }

runs/Dec17_09-25-45_ellis-compute-02.cs.cornell.edu/events.out.tfevents.1734450688.ellis-compute-02.cs.cornell.edu.186689.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57ca8207368c43fc193a9782faab2bbef1c9b404cd80ffe521f99538a86d0816
+size 359