commit files to HF hub

README.md

@@ -1,19 +1,32 @@
 ---
 tags:
 - generated_from_trainer
+datasets:
+- pile-instruct/
 metrics:
 - accuracy
 model-index:
-- name: expert-min-pile-instruct
-  results: []
+- name: layer_4,5,6,7,8
+  results:
+  - task:
+      type: text-generation
+      name: Causal Language Modeling
+    dataset:
+      name: pile-instruct/
+      type: pile-instruct/
+      split: None
+    metrics:
+    - type: accuracy
+      value: 0.20994595912408442
+      name: Accuracy
 ---
 
 <!-- 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. -->
 
-# expert-min-pile-instruct
+# layer_4,5,6,7,8
 
-This model is a fine-tuned version of [P1ayer-1/pythia-deduped-1b-chat-base](https://huggingface.co/P1ayer-1/pythia-deduped-1b-chat-base) on the None dataset.
+This model is a fine-tuned version of [P1ayer-1/pythia-deduped-1b-chat-base](https://huggingface.co/P1ayer-1/pythia-deduped-1b-chat-base) on the pile-instruct/ dataset.
 It achieves the following results on the evaluation set:
 - Loss: 6.9437
 - Accuracy: 0.2099
@@ -65,3 +78,7 @@ The following hyperparameters were used during training:
 - Pytorch 2.0.0+cu117
 - Datasets 2.11.0
 - Tokenizers 0.13.3
+
+
+## Wandb Report
+ https://wandb.ai/ontocord/pythia-1b-deduped-layer-test-min-pile-instruct/runs/6hvfd11h

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 0.08,
+    "eval_accuracy": 0.20994595912408442,
+    "eval_loss": 6.943708896636963,
+    "eval_runtime": 558.0424,
+    "eval_samples": 41323,
+    "eval_samples_per_second": 74.05,
+    "eval_steps_per_second": 1.158,
+    "perplexity": 1036.607765076698,
+    "train_loss": 7.2976521911621095,
+    "train_runtime": 4464.2346,
+    "train_samples": 770564,
+    "train_samples_per_second": 14.336,
+    "train_steps_per_second": 0.224
+}

eval_results.json

@@ -0,0 +1,10 @@
+{
+    "epoch": 0.08,
+    "eval_accuracy": 0.20994595912408442,
+    "eval_loss": 6.943708896636963,
+    "eval_runtime": 558.0424,
+    "eval_samples": 41323,
+    "eval_samples_per_second": 74.05,
+    "eval_steps_per_second": 1.158,
+    "perplexity": 1036.607765076698
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.08,
+    "train_loss": 7.2976521911621095,
+    "train_runtime": 4464.2346,
+    "train_samples": 770564,
+    "train_samples_per_second": 14.336,
+    "train_steps_per_second": 0.224
+}

trainer_state.json

@@ -0,0 +1,370 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.0830556161169423,
+  "global_step": 1000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 9.8e-05,
+      "loss": 9.3581,
+      "step": 20
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 9.6e-05,
+      "loss": 8.7296,
+      "step": 40
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 9.4e-05,
+      "loss": 8.4359,
+      "step": 60
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.200000000000001e-05,
+      "loss": 8.1868,
+      "step": 80
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9e-05,
+      "loss": 8.0439,
+      "step": 100
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.800000000000001e-05,
+      "loss": 7.889,
+      "step": 120
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.6e-05,
+      "loss": 7.8065,
+      "step": 140
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.4e-05,
+      "loss": 7.7108,
+      "step": 160
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.2e-05,
+      "loss": 7.6158,
+      "step": 180
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 8e-05,
+      "loss": 7.6017,
+      "step": 200
+    },
+    {
+      "epoch": 0.02,
+      "eval_accuracy": 0.16046104828397617,
+      "eval_loss": 7.592838764190674,
+      "eval_runtime": 555.8913,
+      "eval_samples_per_second": 74.336,
+      "eval_steps_per_second": 1.162,
+      "step": 200
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 7.800000000000001e-05,
+      "loss": 7.5238,
+      "step": 220
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 7.6e-05,
+      "loss": 7.5017,
+      "step": 240
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 7.4e-05,
+      "loss": 7.4579,
+      "step": 260
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 7.2e-05,
+      "loss": 7.4226,
+      "step": 280
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 7e-05,
+      "loss": 7.3535,
+      "step": 300
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6.800000000000001e-05,
+      "loss": 7.2976,
+      "step": 320
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6.6e-05,
+      "loss": 7.2663,
+      "step": 340
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6.400000000000001e-05,
+      "loss": 7.2552,
+      "step": 360
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6.2e-05,
+      "loss": 7.2456,
+      "step": 380
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6e-05,
+      "loss": 7.1871,
+      "step": 400
+    },
+    {
+      "epoch": 0.03,
+      "eval_accuracy": 0.18468276244618254,
+      "eval_loss": 7.269036769866943,
+      "eval_runtime": 600.1232,
+      "eval_samples_per_second": 68.858,
+      "eval_steps_per_second": 1.076,
+      "step": 400
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.8e-05,
+      "loss": 7.2095,
+      "step": 420
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.6000000000000006e-05,
+      "loss": 7.1898,
+      "step": 440
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.4000000000000005e-05,
+      "loss": 7.2239,
+      "step": 460
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.2000000000000004e-05,
+      "loss": 7.1352,
+      "step": 480
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5e-05,
+      "loss": 7.1078,
+      "step": 500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.8e-05,
+      "loss": 7.1405,
+      "step": 520
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.600000000000001e-05,
+      "loss": 7.0754,
+      "step": 540
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.4000000000000006e-05,
+      "loss": 7.1163,
+      "step": 560
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.2e-05,
+      "loss": 7.0245,
+      "step": 580
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4e-05,
+      "loss": 7.0356,
+      "step": 600
+    },
+    {
+      "epoch": 0.05,
+      "eval_accuracy": 0.19801584131276806,
+      "eval_loss": 7.089658737182617,
+      "eval_runtime": 560.9088,
+      "eval_samples_per_second": 73.672,
+      "eval_steps_per_second": 1.152,
+      "step": 600
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 3.8e-05,
+      "loss": 6.9961,
+      "step": 620
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 3.6e-05,
+      "loss": 7.0128,
+      "step": 640
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 3.4000000000000007e-05,
+      "loss": 7.0096,
+      "step": 660
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.2000000000000005e-05,
+      "loss": 7.0045,
+      "step": 680
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3e-05,
+      "loss": 6.9897,
+      "step": 700
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.8000000000000003e-05,
+      "loss": 6.9639,
+      "step": 720
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.6000000000000002e-05,
+      "loss": 6.9919,
+      "step": 740
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.4e-05,
+      "loss": 6.9351,
+      "step": 760
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.2000000000000003e-05,
+      "loss": 6.8762,
+      "step": 780
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2e-05,
+      "loss": 6.93,
+      "step": 800
+    },
+    {
+      "epoch": 0.07,
+      "eval_accuracy": 0.20642096323588505,
+      "eval_loss": 6.986982822418213,
+      "eval_runtime": 551.4842,
+      "eval_samples_per_second": 74.931,
+      "eval_steps_per_second": 1.171,
+      "step": 800
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.8e-05,
+      "loss": 6.8942,
+      "step": 820
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.6000000000000003e-05,
+      "loss": 6.8943,
+      "step": 840
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.4000000000000001e-05,
+      "loss": 6.9178,
+      "step": 860
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.2e-05,
+      "loss": 6.9095,
+      "step": 880
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1e-05,
+      "loss": 6.9366,
+      "step": 900
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 6.8991,
+      "step": 920
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 6e-06,
+      "loss": 6.8957,
+      "step": 940
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 6.897,
+      "step": 960
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 6.8716,
+      "step": 980
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.0,
+      "loss": 6.9089,
+      "step": 1000
+    },
+    {
+      "epoch": 0.08,
+      "eval_accuracy": 0.20994595912408442,
+      "eval_loss": 6.943708896636963,
+      "eval_runtime": 590.3839,
+      "eval_samples_per_second": 69.993,
+      "eval_steps_per_second": 1.094,
+      "step": 1000
+    },
+    {
+      "epoch": 0.08,
+      "step": 1000,
+      "total_flos": 1.7865923100672e+17,
+      "train_loss": 7.2976521911621095,
+      "train_runtime": 4464.2346,
+      "train_samples_per_second": 14.336,
+      "train_steps_per_second": 0.224
+    }
+  ],
+  "max_steps": 1000,
+  "num_train_epochs": 1,
+  "total_flos": 1.7865923100672e+17,
+  "trial_name": null,
+  "trial_params": null
+}