Model save

README.md

@@ -2,16 +2,12 @@
 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:
-- HuggingFaceH4/ultrachat_200k
+- generator
 model-index:
 - name: gpt2-sft-port
   results: []
@@ -22,7 +18,7 @@ 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 HuggingFaceH4/ultrachat_200k dataset.
+This model is a fine-tuned version of [JackFram/llama-68m](https://huggingface.co/JackFram/llama-68m) on the generator dataset.
 It achieves the following results on the evaluation set:
 - Loss: 2.0739
 

all_results.json

@@ -6,8 +6,8 @@
     "eval_samples_per_second": 1175.543,
     "eval_steps_per_second": 36.76,
     "train_loss": 0.007307469739229727,
-    "train_runtime": 34.7311,
+    "train_runtime": 35.2869,
     "train_samples": 207865,
-    "train_samples_per_second": 16637.979,
-    "train_steps_per_second": 130.027
+    "train_samples_per_second": 16375.952,
+    "train_steps_per_second": 127.98
 }

config.json

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

runs/Apr11_17-34-44_aga39/events.out.tfevents.1712874917.aga39.1781327.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9c7ce4dfe96a7786a3b03c83c18890d2d859cbabce49dc13569fdeb67b80fd0
+size 5953

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 2.0,
     "train_loss": 0.007307469739229727,
-    "train_runtime": 34.7311,
+    "train_runtime": 35.2869,
     "train_samples": 207865,
-    "train_samples_per_second": 16637.979,
-    "train_steps_per_second": 130.027
+    "train_samples_per_second": 16375.952,
+    "train_steps_per_second": 127.98
 }

trainer_state.json

@@ -6325,21 +6325,21 @@
     },
     {
       "epoch": 2.0,
-      "grad_norm": 0.5596599381401034,
+      "grad_norm": 0.559661682911221,
       "learning_rate": 3.6153063554089653e-10,
       "loss": 2.0555,
       "step": 4505
     },
     {
       "epoch": 2.0,
-      "grad_norm": 0.571958581114088,
+      "grad_norm": 0.5719606164762697,
       "learning_rate": 1.0756328901018188e-10,
       "loss": 2.0598,
       "step": 4510
     },
     {
       "epoch": 2.0,
-      "grad_norm": 0.534656420357718,
+      "grad_norm": 0.5346578992785624,
       "learning_rate": 2.987874346827013e-12,
       "loss": 2.054,
       "step": 4515
@@ -6347,9 +6347,9 @@
     {
       "epoch": 2.0,
       "eval_loss": 2.0739028453826904,
-      "eval_runtime": 26.2622,
-      "eval_samples_per_second": 1217.682,
-      "eval_steps_per_second": 38.078,
+      "eval_runtime": 26.6375,
+      "eval_samples_per_second": 1200.527,
+      "eval_steps_per_second": 37.541,
       "step": 4516
     },
     {
@@ -6357,9 +6357,9 @@
       "step": 4516,
       "total_flos": 13637863342080.0,
       "train_loss": 0.007307469739229727,
-      "train_runtime": 34.7311,
-      "train_samples_per_second": 16637.979,
-      "train_steps_per_second": 130.027
+      "train_runtime": 35.2869,
+      "train_samples_per_second": 16375.952,
+      "train_steps_per_second": 127.98
     }
   ],
   "logging_steps": 5,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5178cccce4b851671b9974085a18000b0059f3e1f3f9ae82135b6872a00ff77c
+oid sha256:348f0659098e36c5a4f4c6183168dbfc1b21b9f74e08519031cd01e5095db1a8
 size 6072