huggingartists

README.md

@@ -45,15 +45,15 @@ from datasets import load_dataset
 dataset = load_dataset("huggingartists/zemfira")
 ```
 
-[Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/fzlwlb18/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/3hj4sma8/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 Земфира (Zemfira)'s lyrics.
 
-Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/1uuirro7) for full transparency and reproducibility.
+Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/1v74giz2) for full transparency and reproducibility.
 
-At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/1uuirro7/artifacts) is logged and versioned.
+At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/1v74giz2/artifacts) is logged and versioned.
 
 ## How to use
 

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "gpt2",
+  "_name_or_path": "huggingartists/zemfira",
   "activation_function": "gelu_new",
   "architectures": [
     "GPT2LMHeadModel"

evaluation.txt

@@ -1 +1 @@
-{"eval_loss": 1.7723504304885864, "eval_runtime": 3.5108, "eval_samples_per_second": 22.217, "eval_steps_per_second": 2.848, "epoch": 1.0}
+{"eval_loss": 1.7192894220352173, "eval_runtime": 3.4091, "eval_samples_per_second": 20.827, "eval_steps_per_second": 2.64, "epoch": 2.0}

flax_model.msgpack

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e6459e8282ae04d405f178518877a3524a39a57cf65d9b2b5234a04a2694358c
+oid sha256:4c1f24d5f5d97214ea17828354e30c1cb0ab42893879b8f210bb789754e47b65
 size 497764120

optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dbd6c9f3b863771276c9e277199569516911bf865d02acb367349eb4619a9423
+oid sha256:fb2b4d0752c13f5268ee4193ab28152e257cd319d7febad402156ab94ce40514
 size 995603825

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:15862618e458991c730eb040418ef8bb80dc4a6e9c912d5031be31d5f793d819
+oid sha256:b2c544abb49c657e56c6e5d83148da11b08271b093394d03e86ca288f0a414cf
 size 510403817

rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:83944b113493814ae5c9c809188afdcbb000becae281a11123c189b2d376c1e0
-size 14503
+oid sha256:ed9526fd15362f8845d61d83def08176b3bae047f567421fef1efe555a551faa
+size 14567

scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7b0c69c25efb33387eda1137552cf03b56244a07541d58b5090744f9d9078bbc
+oid sha256:8a232bfd4c68f2b7c5d46c17b697d266203cb8334a94261d8334dd1f3e92002d
 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/zemfira", "tokenizer_class": "GPT2Tokenizer"}

trainer_state.json

@@ -1,78 +1,152 @@
 {
-  "best_metric": 1.7723504304885864,
-  "best_model_checkpoint": "output/zemfira/checkpoint-49",
-  "epoch": 1.0,
-  "global_step": 49,
+  "best_metric": 1.7192894220352173,
+  "best_model_checkpoint": "output/zemfira/checkpoint-100",
+  "epoch": 2.0,
+  "global_step": 100,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.1,
-      "learning_rate": 0.00013370522424493186,
-      "loss": 2.4725,
+      "learning_rate": 3.357522982152468e-06,
+      "loss": 1.9295,
       "step": 5
     },
     {
       "epoch": 0.2,
-      "learning_rate": 0.00012357697446014183,
-      "loss": 2.1691,
+      "learning_rate": 1.3101434185878598e-05,
+      "loss": 1.7553,
       "step": 10
     },
     {
-      "epoch": 0.31,
-      "learning_rate": 0.00010784720288438084,
-      "loss": 1.9883,
+      "epoch": 0.3,
+      "learning_rate": 2.8277931692736335e-05,
+      "loss": 1.9311,
       "step": 15
     },
     {
-      "epoch": 0.41,
-      "learning_rate": 8.811859244288883e-05,
-      "loss": 1.8517,
+      "epoch": 0.4,
+      "learning_rate": 4.74014341858786e-05,
+      "loss": 1.854,
       "step": 20
     },
     {
-      "epoch": 0.51,
-      "learning_rate": 6.640126177858445e-05,
-      "loss": 1.9739,
+      "epoch": 0.5,
+      "learning_rate": 6.859999999999999e-05,
+      "loss": 1.8097,
       "step": 25
     },
     {
-      "epoch": 0.61,
-      "learning_rate": 4.4907957266698305e-05,
-      "loss": 1.9488,
+      "epoch": 0.6,
+      "learning_rate": 8.979856581412138e-05,
+      "loss": 1.9139,
       "step": 30
     },
     {
-      "epoch": 0.71,
-      "learning_rate": 2.5828599592490882e-05,
-      "loss": 1.9148,
+      "epoch": 0.7,
+      "learning_rate": 0.00010892206830726364,
+      "loss": 1.8094,
       "step": 35
     },
     {
-      "epoch": 0.82,
-      "learning_rate": 1.1107156004419736e-05,
-      "loss": 1.9426,
+      "epoch": 0.8,
+      "learning_rate": 0.0001240985658141214,
+      "loss": 1.6989,
       "step": 40
     },
     {
-      "epoch": 0.92,
-      "learning_rate": 2.24357239552258e-06,
-      "loss": 2.0432,
+      "epoch": 0.9,
+      "learning_rate": 0.00013384247701784754,
+      "loss": 1.7624,
       "step": 45
     },
     {
       "epoch": 1.0,
-      "eval_loss": 1.7723504304885864,
-      "eval_runtime": 3.4368,
-      "eval_samples_per_second": 22.696,
-      "eval_steps_per_second": 2.91,
-      "step": 49
+      "learning_rate": 0.0001372,
+      "loss": 1.8189,
+      "step": 50
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 1.7890528440475464,
+      "eval_runtime": 3.3356,
+      "eval_samples_per_second": 21.286,
+      "eval_steps_per_second": 2.698,
+      "step": 50
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.00013384247701784754,
+      "loss": 1.7241,
+      "step": 55
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.00012409856581412136,
+      "loss": 1.7957,
+      "step": 60
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.00010892206830726372,
+      "loss": 1.7292,
+      "step": 65
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 8.979856581412141e-05,
+      "loss": 1.6565,
+      "step": 70
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 6.860000000000001e-05,
+      "loss": 1.6323,
+      "step": 75
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.7401434185878625e-05,
+      "loss": 1.6913,
+      "step": 80
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.827793169273636e-05,
+      "loss": 1.6236,
+      "step": 85
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.310143418587862e-05,
+      "loss": 1.6434,
+      "step": 90
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.3575229821524754e-06,
+      "loss": 1.7407,
+      "step": 95
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 0.0,
+      "loss": 1.7433,
+      "step": 100
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 1.7192894220352173,
+      "eval_runtime": 3.3307,
+      "eval_samples_per_second": 21.317,
+      "eval_steps_per_second": 2.702,
+      "step": 100
     }
   ],
-  "max_steps": 49,
-  "num_train_epochs": 1,
-  "total_flos": 50821300224000.0,
+  "max_steps": 100,
+  "num_train_epochs": 2,
+  "total_flos": 103471644672000.0,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c540afccee3431a837d2b3268ea08f3ec4964cf38f2eac80dc5459e241364e15
+oid sha256:f2e54519d4c930d1b7979099cf1dd3b0fcf472737a97a111bd2ce4bc6adb9b9e
 size 2671