adding model fine-tuned on QA (TAR)

.gitattributes

@@ -26,3 +26,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zstandard filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+eval_nbest_predictions.json filter=lfs diff=lfs merge=lfs -text

all_results.json

@@ -0,0 +1,11 @@
+{
+    "epoch": 4.0,
+    "eval_exact_match": 59.86754966887417,
+    "eval_f1": 76.88812735673795,
+    "eval_samples": 10624,
+    "train_loss": 1.4025724537610833,
+    "train_runtime": 13926.1742,
+    "train_samples": 87747,
+    "train_samples_per_second": 25.203,
+    "train_steps_per_second": 1.575
+}

config.json

@@ -0,0 +1,33 @@
+{
+  "_name_or_path": "CenIA/albert_large_spanish",
+  "architectures": [
+    "AlbertForQuestionAnswering"
+  ],
+  "attention_probs_dropout_prob": 0,
+  "bos_token_id": 2,
+  "classifier_dropout_prob": 0.1,
+  "down_scale_factor": 1,
+  "embedding_size": 128,
+  "eos_token_id": 3,
+  "gap_size": 0,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "inner_group_num": 1,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "albert",
+  "net_structure_type": 0,
+  "num_attention_heads": 16,
+  "num_hidden_groups": 1,
+  "num_hidden_layers": 24,
+  "num_memory_blocks": 0,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "torch_dtype": "float32",
+  "transformers_version": "4.15.0",
+  "type_vocab_size": 2,
+  "vocab_size": 31000
+}

eval_nbest_predictions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1f489b78998a9237dedcb2ac361d0c63fdcbcc5515bc8a749c60930d8d90777
+size 45001705

eval_results.json

@@ -0,0 +1,6 @@
+{
+    "epoch": 4.0,
+    "eval_exact_match": 59.86754966887417,
+    "eval_f1": 76.88812735673795,
+    "eval_samples": 10624
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50a3e8d4919a81a800ccacf93903502786b14d615bba1d75724921eb731a07bc
+size 67071895

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "[CLS]", "eos_token": "[SEP]", "unk_token": "<unk>", "sep_token": "[SEP]", "pad_token": "<pad>", "cls_token": "[CLS]", "mask_token": "[MASK]"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"do_lower_case": true, "remove_space": true, "keep_accents": true, "bos_token": "[CLS]", "eos_token": "[SEP]", "unk_token": "<unk>", "sep_token": "[SEP]", "pad_token": "<pad>", "cls_token": "[CLS]", "mask_token": {"content": "[MASK]", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "CenIA/albert_large_spanish", "tokenizer_class": "AlbertTokenizer"}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 4.0,
+    "train_loss": 1.4025724537610833,
+    "train_runtime": 13926.1742,
+    "train_samples": 87747,
+    "train_samples_per_second": 25.203,
+    "train_steps_per_second": 1.575
+}

trainer_state.json

@@ -0,0 +1,721 @@
+{
+  "best_metric": 76.88812735673795,
+  "best_model_checkpoint": "/data/jcanete/all_results/tar/albeto_large/epochs_4_bs_16_lr_5e-6/checkpoint-15600",
+  "epoch": 3.99990883398669,
+  "global_step": 21936,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.05,
+      "eval_exact_match": 41.7123935666982,
+      "eval_f1": 60.29401087357971,
+      "step": 300
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.887171772428884e-06,
+      "loss": 2.7603,
+      "step": 500
+    },
+    {
+      "epoch": 0.11,
+      "eval_exact_match": 47.88079470198676,
+      "eval_f1": 66.39084020086572,
+      "step": 600
+    },
+    {
+      "epoch": 0.16,
+      "eval_exact_match": 50.52034058656575,
+      "eval_f1": 68.74096558529118,
+      "step": 900
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.773203865791394e-06,
+      "loss": 1.9271,
+      "step": 1000
+    },
+    {
+      "epoch": 0.22,
+      "eval_exact_match": 50.91769157994324,
+      "eval_f1": 70.11766866179605,
+      "step": 1200
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.659235959153902e-06,
+      "loss": 1.8139,
+      "step": 1500
+    },
+    {
+      "epoch": 0.27,
+      "eval_exact_match": 52.469252601702934,
+      "eval_f1": 70.3737951692374,
+      "step": 1500
+    },
+    {
+      "epoch": 0.33,
+      "eval_exact_match": 53.90728476821192,
+      "eval_f1": 71.8131971248393,
+      "step": 1800
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.5452680525164115e-06,
+      "loss": 1.7447,
+      "step": 2000
+    },
+    {
+      "epoch": 0.38,
+      "eval_exact_match": 54.00189214758751,
+      "eval_f1": 72.7403357532424,
+      "step": 2100
+    },
+    {
+      "epoch": 0.44,
+      "eval_exact_match": 54.7682119205298,
+      "eval_f1": 72.50269883199105,
+      "step": 2400
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.431300145878921e-06,
+      "loss": 1.6775,
+      "step": 2500
+    },
+    {
+      "epoch": 0.49,
+      "eval_exact_match": 55.50614947965941,
+      "eval_f1": 73.58281516410962,
+      "step": 2700
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.3173322392414305e-06,
+      "loss": 1.6541,
+      "step": 3000
+    },
+    {
+      "epoch": 0.55,
+      "eval_exact_match": 56.19678334910123,
+      "eval_f1": 74.02233505494875,
+      "step": 3000
+    },
+    {
+      "epoch": 0.6,
+      "eval_exact_match": 56.2724692526017,
+      "eval_f1": 73.79619452732652,
+      "step": 3300
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.203364332603939e-06,
+      "loss": 1.6342,
+      "step": 3500
+    },
+    {
+      "epoch": 0.66,
+      "eval_exact_match": 56.177861873226114,
+      "eval_f1": 73.84756635429511,
+      "step": 3600
+    },
+    {
+      "epoch": 0.71,
+      "eval_exact_match": 57.1996215704825,
+      "eval_f1": 74.85348692600742,
+      "step": 3900
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.089396425966448e-06,
+      "loss": 1.5771,
+      "step": 4000
+    },
+    {
+      "epoch": 0.77,
+      "eval_exact_match": 57.49290444654683,
+      "eval_f1": 75.1912922642055,
+      "step": 4200
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 3.975428519328957e-06,
+      "loss": 1.5866,
+      "step": 4500
+    },
+    {
+      "epoch": 0.82,
+      "eval_exact_match": 57.74834437086093,
+      "eval_f1": 75.25632518139247,
+      "step": 4500
+    },
+    {
+      "epoch": 0.88,
+      "eval_exact_match": 57.47398297067171,
+      "eval_f1": 75.12925316562874,
+      "step": 4800
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.861460612691467e-06,
+      "loss": 1.559,
+      "step": 5000
+    },
+    {
+      "epoch": 0.93,
+      "eval_exact_match": 57.51182592242195,
+      "eval_f1": 74.90745642100545,
+      "step": 5100
+    },
+    {
+      "epoch": 0.98,
+      "eval_exact_match": 57.74834437086093,
+      "eval_f1": 75.43001455522467,
+      "step": 5400
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.7477206418672507e-06,
+      "loss": 1.5324,
+      "step": 5500
+    },
+    {
+      "epoch": 1.04,
+      "eval_exact_match": 58.19299905392621,
+      "eval_f1": 75.72737754816276,
+      "step": 5700
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.63375273522976e-06,
+      "loss": 1.4262,
+      "step": 6000
+    },
+    {
+      "epoch": 1.09,
+      "eval_exact_match": 58.42951750236519,
+      "eval_f1": 75.89034035367936,
+      "step": 6000
+    },
+    {
+      "epoch": 1.15,
+      "eval_exact_match": 57.852412488174075,
+      "eval_f1": 75.61206860863662,
+      "step": 6300
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.5197848285922685e-06,
+      "loss": 1.4304,
+      "step": 6500
+    },
+    {
+      "epoch": 1.2,
+      "eval_exact_match": 58.50520340586566,
+      "eval_f1": 75.88296862635806,
+      "step": 6600
+    },
+    {
+      "epoch": 1.26,
+      "eval_exact_match": 58.27814569536424,
+      "eval_f1": 75.90558506335344,
+      "step": 6900
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.4058169219547776e-06,
+      "loss": 1.4139,
+      "step": 7000
+    },
+    {
+      "epoch": 1.31,
+      "eval_exact_match": 58.666035950804165,
+      "eval_f1": 75.93228571278655,
+      "step": 7200
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.291849015317287e-06,
+      "loss": 1.4231,
+      "step": 7500
+    },
+    {
+      "epoch": 1.37,
+      "eval_exact_match": 58.372753074739826,
+      "eval_f1": 75.62654338728922,
+      "step": 7500
+    },
+    {
+      "epoch": 1.42,
+      "eval_exact_match": 58.98770104068117,
+      "eval_f1": 76.04881705292577,
+      "step": 7800
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.1778811086797962e-06,
+      "loss": 1.4247,
+      "step": 8000
+    },
+    {
+      "epoch": 1.48,
+      "eval_exact_match": 58.760643330179754,
+      "eval_f1": 76.17322174292238,
+      "step": 8100
+    },
+    {
+      "epoch": 1.53,
+      "eval_exact_match": 58.77956480605487,
+      "eval_f1": 76.19180002443377,
+      "step": 8400
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.0639132020423053e-06,
+      "loss": 1.4018,
+      "step": 8500
+    },
+    {
+      "epoch": 1.59,
+      "eval_exact_match": 58.87417218543046,
+      "eval_f1": 76.19402653807724,
+      "step": 8700
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.949945295404814e-06,
+      "loss": 1.4026,
+      "step": 9000
+    },
+    {
+      "epoch": 1.64,
+      "eval_exact_match": 59.205298013245034,
+      "eval_f1": 76.31775571296018,
+      "step": 9000
+    },
+    {
+      "epoch": 1.7,
+      "eval_exact_match": 58.79848628192999,
+      "eval_f1": 76.05081038788326,
+      "step": 9300
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.835977388767323e-06,
+      "loss": 1.3831,
+      "step": 9500
+    },
+    {
+      "epoch": 1.75,
+      "eval_exact_match": 58.902554399243144,
+      "eval_f1": 76.05042386760238,
+      "step": 9600
+    },
+    {
+      "epoch": 1.81,
+      "eval_exact_match": 59.11069063386944,
+      "eval_f1": 76.2790864798194,
+      "step": 9900
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 2.7222374179431077e-06,
+      "loss": 1.3788,
+      "step": 10000
+    },
+    {
+      "epoch": 1.86,
+      "eval_exact_match": 59.13907284768212,
+      "eval_f1": 76.1649200485559,
+      "step": 10200
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 2.608269511305617e-06,
+      "loss": 1.4271,
+      "step": 10500
+    },
+    {
+      "epoch": 1.91,
+      "eval_exact_match": 59.271523178807946,
+      "eval_f1": 76.31194456888919,
+      "step": 10500
+    },
+    {
+      "epoch": 1.97,
+      "eval_exact_match": 59.0728476821192,
+      "eval_f1": 76.3883200778918,
+      "step": 10800
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 2.4943016046681255e-06,
+      "loss": 1.3668,
+      "step": 11000
+    },
+    {
+      "epoch": 2.02,
+      "eval_exact_match": 58.968779564806056,
+      "eval_f1": 76.36879525654898,
+      "step": 11100
+    },
+    {
+      "epoch": 2.08,
+      "eval_exact_match": 59.10122989593188,
+      "eval_f1": 76.61210895681596,
+      "step": 11400
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.3803336980306346e-06,
+      "loss": 1.2558,
+      "step": 11500
+    },
+    {
+      "epoch": 2.13,
+      "eval_exact_match": 59.04446546830653,
+      "eval_f1": 76.39711273020981,
+      "step": 11700
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 2.266593727206419e-06,
+      "loss": 1.2579,
+      "step": 12000
+    },
+    {
+      "epoch": 2.19,
+      "eval_exact_match": 58.94039735099338,
+      "eval_f1": 76.52845917449174,
+      "step": 12000
+    },
+    {
+      "epoch": 2.24,
+      "eval_exact_match": 59.432355723746454,
+      "eval_f1": 76.44161025565445,
+      "step": 12300
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 2.152625820568928e-06,
+      "loss": 1.2599,
+      "step": 12500
+    },
+    {
+      "epoch": 2.3,
+      "eval_exact_match": 59.29990539262062,
+      "eval_f1": 76.46886303091172,
+      "step": 12600
+    },
+    {
+      "epoch": 2.35,
+      "eval_exact_match": 59.508041627246925,
+      "eval_f1": 76.69876166788403,
+      "step": 12900
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 2.038657913931437e-06,
+      "loss": 1.2296,
+      "step": 13000
+    },
+    {
+      "epoch": 2.41,
+      "eval_exact_match": 58.9593188268685,
+      "eval_f1": 76.41311355011096,
+      "step": 13200
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.924690007293946e-06,
+      "loss": 1.2978,
+      "step": 13500
+    },
+    {
+      "epoch": 2.46,
+      "eval_exact_match": 59.39451277199622,
+      "eval_f1": 76.57721571752349,
+      "step": 13500
+    },
+    {
+      "epoch": 2.52,
+      "eval_exact_match": 59.40397350993378,
+      "eval_f1": 76.60409469987792,
+      "step": 13800
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.8109500364697303e-06,
+      "loss": 1.2498,
+      "step": 14000
+    },
+    {
+      "epoch": 2.57,
+      "eval_exact_match": 59.508041627246925,
+      "eval_f1": 76.75425085048117,
+      "step": 14100
+    },
+    {
+      "epoch": 2.63,
+      "eval_exact_match": 59.45127719962157,
+      "eval_f1": 76.45228535462638,
+      "step": 14400
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.6969821298322394e-06,
+      "loss": 1.2848,
+      "step": 14500
+    },
+    {
+      "epoch": 2.68,
+      "eval_exact_match": 59.640491958372756,
+      "eval_f1": 76.65492303716947,
+      "step": 14700
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 1.5830142231947485e-06,
+      "loss": 1.2707,
+      "step": 15000
+    },
+    {
+      "epoch": 2.74,
+      "eval_exact_match": 59.69725638599811,
+      "eval_f1": 76.61197101582836,
+      "step": 15000
+    },
+    {
+      "epoch": 2.79,
+      "eval_exact_match": 59.356669820245976,
+      "eval_f1": 76.54400801502746,
+      "step": 15300
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 1.4690463165572578e-06,
+      "loss": 1.2673,
+      "step": 15500
+    },
+    {
+      "epoch": 2.84,
+      "eval_exact_match": 59.86754966887417,
+      "eval_f1": 76.88812735673795,
+      "step": 15600
+    },
+    {
+      "epoch": 2.9,
+      "eval_exact_match": 59.56480605487228,
+      "eval_f1": 76.78104792015891,
+      "step": 15900
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.3553063457330418e-06,
+      "loss": 1.2817,
+      "step": 16000
+    },
+    {
+      "epoch": 2.95,
+      "eval_exact_match": 59.33774834437086,
+      "eval_f1": 76.65914116498631,
+      "step": 16200
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 1.2413384390955507e-06,
+      "loss": 1.238,
+      "step": 16500
+    },
+    {
+      "epoch": 3.01,
+      "eval_exact_match": 59.75402081362346,
+      "eval_f1": 76.71234598258661,
+      "step": 16500
+    },
+    {
+      "epoch": 3.06,
+      "eval_exact_match": 59.52696310312204,
+      "eval_f1": 76.58652846423291,
+      "step": 16800
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 1.1273705324580598e-06,
+      "loss": 1.1811,
+      "step": 17000
+    },
+    {
+      "epoch": 3.12,
+      "eval_exact_match": 59.120151371807,
+      "eval_f1": 76.43831302372423,
+      "step": 17100
+    },
+    {
+      "epoch": 3.17,
+      "eval_exact_match": 59.31882686849574,
+      "eval_f1": 76.330035615797,
+      "step": 17400
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 1.013630561633844e-06,
+      "loss": 1.1526,
+      "step": 17500
+    },
+    {
+      "epoch": 3.23,
+      "eval_exact_match": 59.46073793755913,
+      "eval_f1": 76.62045789210767,
+      "step": 17700
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 8.996626549963532e-07,
+      "loss": 1.1639,
+      "step": 18000
+    },
+    {
+      "epoch": 3.28,
+      "eval_exact_match": 59.602649006622514,
+      "eval_f1": 76.58821728052384,
+      "step": 18000
+    },
+    {
+      "epoch": 3.34,
+      "eval_exact_match": 59.5364238410596,
+      "eval_f1": 76.48138372732446,
+      "step": 18300
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 7.856947483588622e-07,
+      "loss": 1.1736,
+      "step": 18500
+    },
+    {
+      "epoch": 3.39,
+      "eval_exact_match": 59.39451277199622,
+      "eval_f1": 76.44759593814429,
+      "step": 18600
+    },
+    {
+      "epoch": 3.45,
+      "eval_exact_match": 59.57426679280984,
+      "eval_f1": 76.59464831626804,
+      "step": 18900
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 6.717268417213713e-07,
+      "loss": 1.1616,
+      "step": 19000
+    },
+    {
+      "epoch": 3.5,
+      "eval_exact_match": 59.24314096499527,
+      "eval_f1": 76.46679748121056,
+      "step": 19200
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 5.577589350838805e-07,
+      "loss": 1.1862,
+      "step": 19500
+    },
+    {
+      "epoch": 3.56,
+      "eval_exact_match": 59.356669820245976,
+      "eval_f1": 76.48581357909573,
+      "step": 19500
+    },
+    {
+      "epoch": 3.61,
+      "eval_exact_match": 59.11069063386944,
+      "eval_f1": 76.28951562925242,
+      "step": 19800
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.4379102844638954e-07,
+      "loss": 1.1581,
+      "step": 20000
+    },
+    {
+      "epoch": 3.67,
+      "eval_exact_match": 59.23368022705771,
+      "eval_f1": 76.38635194011843,
+      "step": 20100
+    },
+    {
+      "epoch": 3.72,
+      "eval_exact_match": 59.205298013245034,
+      "eval_f1": 76.5121674711018,
+      "step": 20400
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.298231218088986e-07,
+      "loss": 1.1714,
+      "step": 20500
+    },
+    {
+      "epoch": 3.77,
+      "eval_exact_match": 59.3755912961211,
+      "eval_f1": 76.52009882433762,
+      "step": 20700
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 2.1585521517140777e-07,
+      "loss": 1.1563,
+      "step": 21000
+    },
+    {
+      "epoch": 3.83,
+      "eval_exact_match": 59.33774834437086,
+      "eval_f1": 76.54016429214948,
+      "step": 21000
+    },
+    {
+      "epoch": 3.88,
+      "eval_exact_match": 59.432355723746454,
+      "eval_f1": 76.54804939590761,
+      "step": 21300
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 1.0188730853391687e-07,
+      "loss": 1.1882,
+      "step": 21500
+    },
+    {
+      "epoch": 3.94,
+      "eval_exact_match": 59.34720908230842,
+      "eval_f1": 76.51599845264303,
+      "step": 21600
+    },
+    {
+      "epoch": 3.99,
+      "eval_exact_match": 59.34720908230842,
+      "eval_f1": 76.5156670132116,
+      "step": 21900
+    },
+    {
+      "epoch": 4.0,
+      "step": 21936,
+      "total_flos": 7984258357129344.0,
+      "train_loss": 1.4025724537610833,
+      "train_runtime": 13926.1742,
+      "train_samples_per_second": 25.203,
+      "train_steps_per_second": 1.575
+    }
+  ],
+  "max_steps": 21936,
+  "num_train_epochs": 4,
+  "total_flos": 7984258357129344.0,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b7a4427157b8ec0c345d985bb0ac9d9aaf2a0b18df2cdf5525149c6a8d97d39d
+size 2991