End of training

README.md

@@ -2,12 +2,16 @@
 license: apache-2.0
 base_model: JackFram/llama-68m
 tags:
+- alignment-handbook
+- trl
+- sft
+- generated_from_trainer
 - trl
 - sft
 - alignment-handbook
 - generated_from_trainer
 datasets:
-- generator
+- HuggingFaceH4/ultrachat_200k
 model-index:
 - name: gpt2-sft-port
   results: []
@@ -18,7 +22,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # gpt2-sft-port
 
-This model is a fine-tuned version of [JackFram/llama-68m](https://huggingface.co/JackFram/llama-68m) on the generator dataset.
+This model is a fine-tuned version of [JackFram/llama-68m](https://huggingface.co/JackFram/llama-68m) on the HuggingFaceH4/ultrachat_200k dataset.
 It achieves the following results on the evaluation set:
 - Loss: 2.1067
 

all_results.json

@@ -1,10 +1,10 @@
 {
     "epoch": 2.0,
-    "eval_loss": 2.106668472290039,
-    "eval_runtime": 22.5861,
+    "eval_loss": 2.106666326522827,
+    "eval_runtime": 22.5513,
     "eval_samples": 23110,
-    "eval_samples_per_second": 1415.872,
-    "eval_steps_per_second": 22.138,
+    "eval_samples_per_second": 1418.056,
+    "eval_steps_per_second": 22.172,
     "train_loss": 2.1773925889794863,
     "train_runtime": 649.8909,
     "train_samples": 207865,

config.json

@@ -24,6 +24,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.38.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 2.0,
-    "eval_loss": 2.106668472290039,
-    "eval_runtime": 22.5861,
+    "eval_loss": 2.106666326522827,
+    "eval_runtime": 22.5513,
     "eval_samples": 23110,
-    "eval_samples_per_second": 1415.872,
-    "eval_steps_per_second": 22.138
+    "eval_samples_per_second": 1418.056,
+    "eval_steps_per_second": 22.172
 }

runs/Apr24_18-48-37_aga39/events.out.tfevents.1714003233.aga39.662109.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a764822ea4b8256c21d3918af515439f78b3312854bbad378b82ab346b5eb8d
+size 359