Model save

README.md

@@ -1,15 +1,10 @@
 ---
-base_model: data/gemma-2b
 tags:
-- alignment-handbook
-- trl
-- sft
-- generated_from_trainer
 - trl
 - sft
 - generated_from_trainer
 datasets:
-- argilla/dpo-mix-7k
+- generator
 model-index:
 - name: zephyr-2b-gemma-dft
   results: []
@@ -18,12 +13,10 @@ model-index:
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 should probably proofread and complete it, then remove this comment. -->
 
-[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](None)
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://zebra.wandb.io/cto/distillm/runs/vxz2pnbq)
 # zephyr-2b-gemma-dft
 
-This model is a fine-tuned version of [data/gemma-2b](https://huggingface.co/data/gemma-2b) on the argilla/dpo-mix-7k dataset.
-It achieves the following results on the evaluation set:
-- Loss: 0.0001
+This model was trained from scratch on the generator dataset.
 
 ## Model description
 

all_results.json

@@ -1,13 +1,8 @@
 {
     "epoch": 0,
-    "eval_loss": 5.277253148960881e-05,
-    "eval_runtime": 76.5403,
-    "eval_samples": 750,
-    "eval_samples_per_second": 12.686,
-    "eval_steps_per_second": 1.594,
     "total_flos": 0,
     "train_loss": 0.0,
-    "train_runtime": 4.808,
+    "train_runtime": 5.1094,
     "train_samples": 6750,
     "train_samples_per_second": 0.0,
     "train_steps_per_second": 0.0

config.json

@@ -24,6 +24,6 @@
   "rope_theta": 10000.0,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.42.4",
-  "use_cache": true,
+  "use_cache": false,
   "vocab_size": 256000
 }

runs/Jul23_16-11-48_ale-distillm-8-0-0/events.out.tfevents.1721747524.ale-distillm-8-0-0.69762.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6c935c52bd5493f2494a3ddfe9b4b9ee1dd74408242cad4fd703bd7cce3cf27
+size 5428

train_results.json

@@ -2,7 +2,7 @@
     "epoch": 0,
     "total_flos": 0,
     "train_loss": 0.0,
-    "train_runtime": 4.808,
+    "train_runtime": 5.1094,
     "train_samples": 6750,
     "train_samples_per_second": 0.0,
     "train_steps_per_second": 0.0

trainer_state.json

@@ -13,7 +13,7 @@
       "step": 0,
       "total_flos": 0,
       "train_loss": 0.0,
-      "train_runtime": 4.808,
+      "train_runtime": 5.1094,
       "train_samples_per_second": 0.0,
       "train_steps_per_second": 0.0
     }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:43a0172f6f69bee07075ae9f14129953a7c9c530502a5beeaa13f998fa945ac5
+oid sha256:c1954d4b8d21dac7c6689bc226f36245765ac0d0c5983123291271c68ca5ad82
 size 5304