huggingartists

README.md

@@ -45,15 +45,15 @@ from datasets import load_dataset
 dataset = load_dataset("huggingartists/elton-john")
 ```
 
-[Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/4l9f8jif/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
+[Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/188xpm2n/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
 
 ## Training procedure
 
 The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on Elton John's lyrics.
 
-Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/9glyfrcq) for full transparency and reproducibility.
+Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/1rgstntu) for full transparency and reproducibility.
 
-At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/9glyfrcq/artifacts) is logged and versioned.
+At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/1rgstntu/artifacts) is logged and versioned.
 
 ## How to use
 

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "gpt2",
+  "_name_or_path": "elton-john",
   "activation_function": "gelu_new",
   "architectures": [
     "GPT2LMHeadModel"
@@ -18,7 +18,9 @@
   "n_inner": null,
   "n_layer": 12,
   "n_positions": 1024,
+  "reorder_and_upcast_attn": false,
   "resid_pdrop": 0.1,
+  "scale_attn_by_inverse_layer_idx": false,
   "scale_attn_weights": true,
   "summary_activation": null,
   "summary_first_dropout": 0.1,
@@ -35,7 +37,7 @@
     }
   },
   "torch_dtype": "float32",
-  "transformers_version": "4.10.3",
+  "transformers_version": "4.19.2",
   "use_cache": true,
   "vocab_size": 50257
 }

evaluation.txt

@@ -1 +1 @@
-{"eval_loss": 3.0393269062042236, "eval_runtime": 9.0147, "eval_samples_per_second": 22.852, "eval_steps_per_second": 2.884, "epoch": 1.0}
+{"eval_loss": 2.6964175701141357, "eval_runtime": 2.7971, "eval_samples_per_second": 81.512, "eval_steps_per_second": 10.368, "epoch": 2.0}

flax_model.msgpack

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:80c5335361f4b5826d3c8dcbf64dc173264028aa9a8ed9f51f27f3025014f6dc
+oid sha256:b667c3739abcdabedf828ba5c341005b25b86b08dc2494101a9f9c34e7b23379
 size 497764120

optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fb57694a5fd1f450d99621dea90750bd8c71a315ae39a7ac81bf32c09f7c1c07
-size 995603825
+oid sha256:086dbb4d18bbc687bfdc82ae0265198008c27dfa34ad39ee138d2269363dc21f
+size 995604017

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cc049c7460156279f44ea714b704592cadbf8d4a7d08b8b2a7e4e581447462f9
-size 510403817
+oid sha256:be5b0fb3f09fe35cd579daab9e9e17abf545cd40b3ef49b348439a6496cc8651
+size 510396521

rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2dcaf38b2ba417b9e7c855abc82748c9660bcba4c77392d9424bf18c41d62c2a
+oid sha256:abcffe8f8e425f012a63f6022fd9d4d4c1c45516cf3d4653327a424d11b936f3
 size 14503

scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bb71b8d9941063c635df32e061d21e647151549cf90fc9dd70cfed3b990a34f9
+oid sha256:bc07e42f7b3eb4494cdbc9bcb8911780143ce3b5a9cba5c3f2005a37846c6c2b
 size 623

tokenizer_config.json

@@ -1 +1 @@
-{"unk_token": "<|endoftext|>", "bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "add_prefix_space": false, "model_max_length": 1024, "special_tokens_map_file": null, "name_or_path": "gpt2", "tokenizer_class": "GPT2Tokenizer"}
+{"unk_token": "<|endoftext|>", "bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "add_prefix_space": false, "model_max_length": 1024, "special_tokens_map_file": null, "name_or_path": "huggingartists/elton-john", "tokenizer_class": "GPT2Tokenizer"}

trainer_state.json

@@ -1,8 +1,8 @@
 {
-  "best_metric": 3.0393269062042236,
-  "best_model_checkpoint": "output/elton-john/checkpoint-161",
-  "epoch": 1.0,
-  "global_step": 161,
+  "best_metric": 2.6964175701141357,
+  "best_model_checkpoint": "output/elton-john/checkpoint-316",
+  "epoch": 2.0,
+  "global_step": 316,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -206,11 +206,205 @@
       "eval_samples_per_second": 23.351,
       "eval_steps_per_second": 2.947,
       "step": 161
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 6.633985999863723e-07,
+      "loss": 3.0203,
+      "step": 165
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.94348327322787e-06,
+      "loss": 2.7138,
+      "step": 170
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.881847639061565e-06,
+      "loss": 3.1889,
+      "step": 175
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 6.459348994545111e-06,
+      "loss": 3.1967,
+      "step": 180
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 9.650532712235251e-06,
+      "loss": 2.5474,
+      "step": 185
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.3423883622425528e-05,
+      "loss": 3.0815,
+      "step": 190
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.7742137247474782e-05,
+      "loss": 2.8284,
+      "step": 195
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.256264781457008e-05,
+      "loss": 2.9571,
+      "step": 200
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.783780941254068e-05,
+      "loss": 2.8924,
+      "step": 205
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.3515526133505226e-05,
+      "loss": 3.1057,
+      "step": 210
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.953972655637492e-05,
+      "loss": 2.7295,
+      "step": 215
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.585091749132818e-05,
+      "loss": 2.9389,
+      "step": 220
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 5.238677151664272e-05,
+      "loss": 2.9111,
+      "step": 225
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 5.908274250554988e-05,
+      "loss": 2.9365,
+      "step": 230
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 6.587270306435089e-05,
+      "loss": 2.9594,
+      "step": 235
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 7.268959758664963e-05,
+      "loss": 3.0584,
+      "step": 240
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 7.946610447433497e-05,
+      "loss": 2.8725,
+      "step": 245
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 8.613530098541655e-05,
+      "loss": 2.8485,
+      "step": 250
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 9.263132414287828e-05,
+      "loss": 3.0395,
+      "step": 255
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 9.889002117761157e-05,
+      "loss": 2.9287,
+      "step": 260
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.0001048495830818497,
+      "loss": 3.1458,
+      "step": 265
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.000110451155016322,
+      "loss": 2.9484,
+      "step": 270
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.00011563941754292792,
+      "loss": 2.8225,
+      "step": 275
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.00012036313294285455,
+      "loss": 2.9005,
+      "step": 280
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.00012457565122486245,
+      "loss": 2.915,
+      "step": 285
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.00012823537082655537,
+      "loss": 2.7606,
+      "step": 290
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.00013130614945888676,
+      "loss": 3.1073,
+      "step": 295
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.0001337576610365259,
+      "loss": 2.7606,
+      "step": 300
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.0001355656951691469,
+      "loss": 2.9132,
+      "step": 305
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 0.00013671239625595536,
+      "loss": 3.0598,
+      "step": 310
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 0.0001371864398222265,
+      "loss": 2.8386,
+      "step": 315
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 2.6964175701141357,
+      "eval_runtime": 2.7786,
+      "eval_samples_per_second": 82.056,
+      "eval_steps_per_second": 10.437,
+      "step": 316
     }
   ],
-  "max_steps": 161,
-  "num_train_epochs": 1,
-  "total_flos": 167880130560000.0,
+  "max_steps": 316,
+  "num_train_epochs": 2,
+  "total_flos": 329750544384000.0,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e4c9d72369b89deab6316d05e8deee4071405452ec742851b48a28764b370d00
-size 2671
+oid sha256:4762149b1593c73d8a775519e9ab404b0d7158c9fb60f78e7ba0b4c58ea04c5e
+size 3247