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"/>](https://zebra.wandb.io/cto/distillm/runs/kje9xugl)
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](None)
 # 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: 0.0000
+This model was trained from scratch on the generator dataset.
 
 ## Model description
 
@@ -51,13 +44,10 @@ The following hyperparameters were used during training:
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 1
+- num_epochs: 0
 
 ### Training results
 
-| Training Loss | Epoch  | Step | Validation Loss |
-|:-------------:|:------:|:----:|:---------------:|
-| 0.0           | 0.9982 | 270  | 0.0000          |
 
 
 ### Framework versions

all_results.json

@@ -1,14 +1,9 @@
 {
-    "epoch": 0.9981515711645101,
-    "eval_loss": 1.3726342331210617e-05,
-    "eval_runtime": 50.9693,
-    "eval_samples": 750,
-    "eval_samples_per_second": 19.051,
-    "eval_steps_per_second": 2.394,
-    "total_flos": 5.260333472022528e+16,
-    "train_loss": 1.8035727003330572e-05,
-    "train_runtime": 2689.0507,
+    "epoch": 0,
+    "total_flos": 0,
+    "train_loss": 0.0,
+    "train_runtime": 4.808,
     "train_samples": 6750,
-    "train_samples_per_second": 3.217,
-    "train_steps_per_second": 0.1
+    "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
 }

model-00001-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:166c2cc2ac8f29542b11f7812cc6cdded64345b787b9b74e50c194790584c068
+oid sha256:98caad07f1e1d17fede7734906d6881d4cb231e1d0cd2ebcf7fc8a6317eee4d6
 size 4945242264

model-00002-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:403d33ecb881a9ae12700a261993eb4f1ad36a361fbb628ffaa31bf3ff3456e1
+oid sha256:3342a2f7db857bc9873c699eef1416eb2b320603e5e042a6f7736689b26382cb
 size 67121608

runs/Jul23_16-03-42_ale-distillm-8-0-0/events.out.tfevents.1721747038.ale-distillm-8-0-0.67747.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e00a4e18a9b4b510ffada73fab88f7ab13cc0ec51603704a110c340d5f9e7f99
+size 5428

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 0.9981515711645101,
-    "total_flos": 5.260333472022528e+16,
-    "train_loss": 1.8035727003330572e-05,
-    "train_runtime": 2689.0507,
+    "epoch": 0,
+    "total_flos": 0,
+    "train_loss": 0.0,
+    "train_runtime": 4.808,
     "train_samples": 6750,
-    "train_samples_per_second": 3.217,
-    "train_steps_per_second": 0.1
+    "train_samples_per_second": 0.0,
+    "train_steps_per_second": 0.0
 }

trainer_state.json

@@ -1,413 +1,27 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.9981515711645101,
+  "epoch": 0,
   "eval_steps": 500,
-  "global_step": 270,
+  "global_step": 0,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.018484288354898338,
-      "grad_norm": 0.00118255615234375,
-      "learning_rate": 3.7037037037037037e-06,
-      "loss": 0.0001,
-      "step": 5
-    },
-    {
-      "epoch": 0.036968576709796676,
-      "grad_norm": 0.000579833984375,
-      "learning_rate": 7.4074074074074075e-06,
-      "loss": 0.0001,
-      "step": 10
-    },
-    {
-      "epoch": 0.05545286506469501,
-      "grad_norm": 0.000614166259765625,
-      "learning_rate": 1.1111111111111113e-05,
-      "loss": 0.0,
-      "step": 15
-    },
-    {
-      "epoch": 0.07393715341959335,
-      "grad_norm": 0.000850677490234375,
-      "learning_rate": 1.4814814814814815e-05,
-      "loss": 0.0,
-      "step": 20
-    },
-    {
-      "epoch": 0.09242144177449169,
-      "grad_norm": 0.000446319580078125,
-      "learning_rate": 1.851851851851852e-05,
-      "loss": 0.0,
-      "step": 25
-    },
-    {
-      "epoch": 0.11090573012939002,
-      "grad_norm": 0.00075531005859375,
-      "learning_rate": 1.9992479525042305e-05,
-      "loss": 0.0,
-      "step": 30
-    },
-    {
-      "epoch": 0.12939001848428835,
-      "grad_norm": 0.0003108978271484375,
-      "learning_rate": 1.9946562024066018e-05,
-      "loss": 0.0,
-      "step": 35
-    },
-    {
-      "epoch": 0.1478743068391867,
-      "grad_norm": 0.0003833770751953125,
-      "learning_rate": 1.9859096633447965e-05,
-      "loss": 0.0,
-      "step": 40
-    },
-    {
-      "epoch": 0.16635859519408502,
-      "grad_norm": 0.000499725341796875,
-      "learning_rate": 1.973044870579824e-05,
-      "loss": 0.0,
-      "step": 45
-    },
-    {
-      "epoch": 0.18484288354898337,
-      "grad_norm": 0.000118255615234375,
-      "learning_rate": 1.95611556177388e-05,
-      "loss": 0.0,
-      "step": 50
-    },
-    {
-      "epoch": 0.2033271719038817,
-      "grad_norm": 7.724761962890625e-05,
-      "learning_rate": 1.93519245252219e-05,
-      "loss": 0.0,
-      "step": 55
-    },
-    {
-      "epoch": 0.22181146025878004,
-      "grad_norm": 0.0002689361572265625,
-      "learning_rate": 1.9103629409661468e-05,
-      "loss": 0.0,
-      "step": 60
-    },
-    {
-      "epoch": 0.24029574861367836,
-      "grad_norm": 8.821487426757812e-05,
-      "learning_rate": 1.881730742721608e-05,
-      "loss": 0.0,
-      "step": 65
-    },
-    {
-      "epoch": 0.2587800369685767,
-      "grad_norm": 0.00016307830810546875,
-      "learning_rate": 1.8494154576472976e-05,
-      "loss": 0.0,
-      "step": 70
-    },
-    {
-      "epoch": 0.27726432532347506,
-      "grad_norm": 9.441375732421875e-05,
-      "learning_rate": 1.8135520702629677e-05,
-      "loss": 0.0,
-      "step": 75
-    },
-    {
-      "epoch": 0.2957486136783734,
-      "grad_norm": 0.0002689361572265625,
-      "learning_rate": 1.7742903859041324e-05,
-      "loss": 0.0,
-      "step": 80
-    },
-    {
-      "epoch": 0.3142329020332717,
-      "grad_norm": 5.793571472167969e-05,
-      "learning_rate": 1.7317944049686125e-05,
-      "loss": 0.0,
-      "step": 85
-    },
-    {
-      "epoch": 0.33271719038817005,
-      "grad_norm": 0.00018596649169921875,
-      "learning_rate": 1.686241637868734e-05,
-      "loss": 0.0,
-      "step": 90
-    },
-    {
-      "epoch": 0.3512014787430684,
-      "grad_norm": 0.00012111663818359375,
-      "learning_rate": 1.637822363550706e-05,
-      "loss": 0.0,
-      "step": 95
-    },
-    {
-      "epoch": 0.36968576709796674,
-      "grad_norm": 0.00013446807861328125,
-      "learning_rate": 1.586738834678418e-05,
-      "loss": 0.0,
-      "step": 100
-    },
-    {
-      "epoch": 0.38817005545286504,
-      "grad_norm": 6.246566772460938e-05,
-      "learning_rate": 1.5332044328016916e-05,
-      "loss": 0.0,
-      "step": 105
-    },
-    {
-      "epoch": 0.4066543438077634,
-      "grad_norm": 0.00013828277587890625,
-      "learning_rate": 1.4774427770379492e-05,
-      "loss": 0.0,
-      "step": 110
-    },
-    {
-      "epoch": 0.42513863216266173,
-      "grad_norm": 0.0001544952392578125,
-      "learning_rate": 1.4196867899904292e-05,
-      "loss": 0.0,
-      "step": 115
-    },
-    {
-      "epoch": 0.4436229205175601,
-      "grad_norm": 0.00011968612670898438,
-      "learning_rate": 1.3601777248047105e-05,
-      "loss": 0.0,
-      "step": 120
-    },
-    {
-      "epoch": 0.46210720887245843,
-      "grad_norm": 6.532669067382812e-05,
-      "learning_rate": 1.2991641574276419e-05,
-      "loss": 0.0,
-      "step": 125
-    },
-    {
-      "epoch": 0.4805914972273567,
-      "grad_norm": 9.965896606445312e-05,
-      "learning_rate": 1.2369009482781191e-05,
-      "loss": 0.0,
-      "step": 130
-    },
-    {
-      "epoch": 0.49907578558225507,
-      "grad_norm": 5.698204040527344e-05,
-      "learning_rate": 1.1736481776669307e-05,
-      "loss": 0.0,
-      "step": 135
-    },
-    {
-      "epoch": 0.5175600739371534,
-      "grad_norm": 4.124641418457031e-05,
-      "learning_rate": 1.1096700594125318e-05,
-      "loss": 0.0,
-      "step": 140
-    },
-    {
-      "epoch": 0.5360443622920518,
-      "grad_norm": 9.5367431640625e-05,
-      "learning_rate": 1.0452338371907065e-05,
-      "loss": 0.0,
-      "step": 145
-    },
-    {
-      "epoch": 0.5545286506469501,
-      "grad_norm": 5.698204040527344e-05,
-      "learning_rate": 9.806086682281759e-06,
-      "loss": 0.0,
-      "step": 150
-    },
-    {
-      "epoch": 0.5730129390018485,
-      "grad_norm": 9.441375732421875e-05,
-      "learning_rate": 9.160644990030932e-06,
-      "loss": 0.0,
-      "step": 155
-    },
-    {
-      "epoch": 0.5914972273567468,
-      "grad_norm": 0.00011777877807617188,
-      "learning_rate": 8.518709376487515e-06,
-      "loss": 0.0,
-      "step": 160
-    },
-    {
-      "epoch": 0.609981515711645,
-      "grad_norm": 0.0001583099365234375,
-      "learning_rate": 7.882961277705897e-06,
-      "loss": 0.0,
-      "step": 165
-    },
-    {
-      "epoch": 0.6284658040665434,
-      "grad_norm": 0.000141143798828125,
-      "learning_rate": 7.256056283806987e-06,
-      "loss": 0.0,
-      "step": 170
-    },
-    {
-      "epoch": 0.6469500924214417,
-      "grad_norm": 0.0001697540283203125,
-      "learning_rate": 6.640613046284581e-06,
-      "loss": 0.0,
-      "step": 175
-    },
-    {
-      "epoch": 0.6654343807763401,
-      "grad_norm": 0.000431060791015625,
-      "learning_rate": 6.039202339608432e-06,
-      "loss": 0.0,
-      "step": 180
-    },
-    {
-      "epoch": 0.6839186691312384,
-      "grad_norm": 8.106231689453125e-05,
-      "learning_rate": 5.454336322814995e-06,
-      "loss": 0.0,
-      "step": 185
-    },
-    {
-      "epoch": 0.7024029574861368,
-      "grad_norm": 9.918212890625e-05,
-      "learning_rate": 4.888458045941269e-06,
-      "loss": 0.0,
-      "step": 190
-    },
-    {
-      "epoch": 0.7208872458410351,
-      "grad_norm": 4.76837158203125e-05,
-      "learning_rate": 4.343931245134616e-06,
-      "loss": 0.0,
-      "step": 195
-    },
-    {
-      "epoch": 0.7393715341959335,
-      "grad_norm": 0.00012493133544921875,
-      "learning_rate": 3.823030469065431e-06,
-      "loss": 0.0,
-      "step": 200
-    },
-    {
-      "epoch": 0.7578558225508318,
-      "grad_norm": 5.340576171875e-05,
-      "learning_rate": 3.3279315778858034e-06,
-      "loss": 0.0,
-      "step": 205
-    },
-    {
-      "epoch": 0.7763401109057301,
-      "grad_norm": 7.05718994140625e-05,
-      "learning_rate": 2.8607026544210115e-06,
-      "loss": 0.0,
-      "step": 210
-    },
-    {
-      "epoch": 0.7948243992606284,
-      "grad_norm": 0.0002117156982421875,
-      "learning_rate": 2.423295365558821e-06,
-      "loss": 0.0,
-      "step": 215
-    },
-    {
-      "epoch": 0.8133086876155268,
-      "grad_norm": 4.38690185546875e-05,
-      "learning_rate": 2.01753680992107e-06,
-      "loss": 0.0,
-      "step": 220
-    },
-    {
-      "epoch": 0.8317929759704251,
-      "grad_norm": 0.00010395050048828125,
-      "learning_rate": 1.6451218858706374e-06,
-      "loss": 0.0,
-      "step": 225
-    },
-    {
-      "epoch": 0.8502772643253235,
-      "grad_norm": 0.00018978118896484375,
-      "learning_rate": 1.307606211733522e-06,
-      "loss": 0.0,
-      "step": 230
-    },
-    {
-      "epoch": 0.8687615526802218,
-      "grad_norm": 5.53131103515625e-05,
-      "learning_rate": 1.0063996278090704e-06,
-      "loss": 0.0,
-      "step": 235
-    },
-    {
-      "epoch": 0.8872458410351202,
-      "grad_norm": 6.389617919921875e-05,
-      "learning_rate": 7.427603073110967e-07,
-      "loss": 0.0,
-      "step": 240
-    },
-    {
-      "epoch": 0.9057301293900185,
-      "grad_norm": 0.00011205673217773438,
-      "learning_rate": 5.177895008392353e-07,
-      "loss": 0.0,
-      "step": 245
-    },
-    {
-      "epoch": 0.9242144177449169,
-      "grad_norm": 0.000293731689453125,
-      "learning_rate": 3.3242693633337986e-07,
-      "loss": 0.0,
-      "step": 250
-    },
-    {
-      "epoch": 0.9426987060998152,
-      "grad_norm": 0.000179290771484375,
-      "learning_rate": 1.874468937261531e-07,
-      "loss": 0.0,
-      "step": 255
-    },
-    {
-      "epoch": 0.9611829944547134,
-      "grad_norm": 0.002349853515625,
-      "learning_rate": 8.345497068998897e-08,
-      "loss": 0.0,
-      "step": 260
-    },
-    {
-      "epoch": 0.9796672828096118,
-      "grad_norm": 6.437301635742188e-05,
-      "learning_rate": 2.088555298867978e-08,
-      "loss": 0.0,
-      "step": 265
-    },
-    {
-      "epoch": 0.9981515711645101,
-      "grad_norm": 9.1552734375e-05,
-      "learning_rate": 0.0,
-      "loss": 0.0,
-      "step": 270
-    },
-    {
-      "epoch": 0.9981515711645101,
-      "eval_loss": 1.3726342331210617e-05,
-      "eval_runtime": 52.2001,
-      "eval_samples_per_second": 18.601,
-      "eval_steps_per_second": 2.337,
-      "step": 270
-    },
-    {
-      "epoch": 0.9981515711645101,
-      "step": 270,
-      "total_flos": 5.260333472022528e+16,
-      "train_loss": 1.8035727003330572e-05,
-      "train_runtime": 2689.0507,
-      "train_samples_per_second": 3.217,
-      "train_steps_per_second": 0.1
+      "epoch": 0,
+      "step": 0,
+      "total_flos": 0,
+      "train_loss": 0.0,
+      "train_runtime": 4.808,
+      "train_samples_per_second": 0.0,
+      "train_steps_per_second": 0.0
     }
   ],
   "logging_steps": 5,
-  "max_steps": 270,
+  "max_steps": 0,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 1,
+  "num_train_epochs": 0,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -421,7 +35,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 5.260333472022528e+16,
+  "total_flos": 0,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:48db1adcd7121cc0952bac83481593034c426cd3d56b3e7a70068bc17b1bad10
+oid sha256:43a0172f6f69bee07075ae9f14129953a7c9c530502a5beeaa13f998fa945ac5
 size 5304