Gaussian 512 fine-tuned PAWS-X es with acc 0.8965

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 5.0,
+    "eval_accuracy": 0.8965,
+    "eval_loss": 0.5265980958938599,
+    "eval_runtime": 21.0936,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 94.815,
+    "eval_steps_per_second": 5.926,
+    "train_loss": 0.1811776852978326,
+    "train_runtime": 8418.9007,
+    "train_samples": 49401,
+    "train_samples_per_second": 29.339,
+    "train_steps_per_second": 1.834
+}

config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "bertin-project/bertin-base-gaussian-exp-512seqlen",
+  "architectures": [
+    "RobertaForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "eos_token_id": 2,
+  "finetuning_task": "paws-x",
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "roberta",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "torch_dtype": "float32",
+  "transformers_version": "4.9.0.dev0",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 50265
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 5.0,
+    "eval_accuracy": 0.8965,
+    "eval_loss": 0.5265980958938599,
+    "eval_runtime": 21.0936,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 94.815,
+    "eval_steps_per_second": 5.926
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f33890dc060a5e2908bbd3d4fe506d7ddc8194086f80f92d841e8c985f6f7b73
+size 498674093

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": false}}

tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "add_prefix_space": false, "errors": "replace", "sep_token": "</s>", "cls_token": "<s>", "pad_token": "<pad>", "mask_token": "<mask>", "special_tokens_map_file": null, "name_or_path": "bertin-project/bertin-base-gaussian-exp-512seqlen", "tokenizer_class": "RobertaTokenizer"}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 5.0,
+    "train_loss": 0.1811776852978326,
+    "train_runtime": 8418.9007,
+    "train_samples": 49401,
+    "train_samples_per_second": 29.339,
+    "train_steps_per_second": 1.834
+}

trainer_state.json

@@ -0,0 +1,205 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 5.0,
+  "global_step": 15440,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.8380829015544046e-05,
+      "loss": 0.5771,
+      "step": 500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.676165803108808e-05,
+      "loss": 0.3961,
+      "step": 1000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.5142487046632126e-05,
+      "loss": 0.3378,
+      "step": 1500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.352331606217617e-05,
+      "loss": 0.3113,
+      "step": 2000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.190414507772021e-05,
+      "loss": 0.2978,
+      "step": 2500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.028497409326425e-05,
+      "loss": 0.269,
+      "step": 3000
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.8665803108808294e-05,
+      "loss": 0.2396,
+      "step": 3500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.704663212435233e-05,
+      "loss": 0.215,
+      "step": 4000
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.5427461139896374e-05,
+      "loss": 0.215,
+      "step": 4500
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.380829015544041e-05,
+      "loss": 0.1969,
+      "step": 5000
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.2189119170984454e-05,
+      "loss": 0.1983,
+      "step": 5500
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.05699481865285e-05,
+      "loss": 0.2113,
+      "step": 6000
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.8950777202072538e-05,
+      "loss": 0.1659,
+      "step": 6500
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 2.7331606217616585e-05,
+      "loss": 0.1544,
+      "step": 7000
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 2.5712435233160625e-05,
+      "loss": 0.1517,
+      "step": 7500
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.4093264248704665e-05,
+      "loss": 0.156,
+      "step": 8000
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 2.2474093264248706e-05,
+      "loss": 0.1547,
+      "step": 8500
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 2.0854922279792746e-05,
+      "loss": 0.1331,
+      "step": 9000
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 1.9235751295336786e-05,
+      "loss": 0.1278,
+      "step": 9500
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 1.761658031088083e-05,
+      "loss": 0.1125,
+      "step": 10000
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 1.5997409326424873e-05,
+      "loss": 0.1148,
+      "step": 10500
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 1.4378238341968913e-05,
+      "loss": 0.1123,
+      "step": 11000
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 1.2759067357512955e-05,
+      "loss": 0.106,
+      "step": 11500
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 1.1139896373056995e-05,
+      "loss": 0.1053,
+      "step": 12000
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 9.520725388601037e-06,
+      "loss": 0.0906,
+      "step": 12500
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 7.901554404145079e-06,
+      "loss": 0.0792,
+      "step": 13000
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 6.282383419689119e-06,
+      "loss": 0.0726,
+      "step": 13500
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 4.663212435233161e-06,
+      "loss": 0.0781,
+      "step": 14000
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 3.044041450777202e-06,
+      "loss": 0.0697,
+      "step": 14500
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 1.4248704663212437e-06,
+      "loss": 0.075,
+      "step": 15000
+    },
+    {
+      "epoch": 5.0,
+      "step": 15440,
+      "total_flos": 6.49897462291968e+16,
+      "train_loss": 0.1811776852978326,
+      "train_runtime": 8418.9007,
+      "train_samples_per_second": 29.339,
+      "train_steps_per_second": 1.834
+    }
+  ],
+  "max_steps": 15440,
+  "num_train_epochs": 5,
+  "total_flos": 6.49897462291968e+16,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:576eca8f4ddd686c360a049868136f08fcaf73a619b5fc176fba347c70891c2e
+size 2735