Push to HF hub

.gitattributes

@@ -31,3 +31,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,13 @@
+checkpoint-15000/
+events.out.tfevents.1663038340.vesna.703613.2
+events.out.tfevents.1662992784.vesna.703613.0
+1662992784.275189/
+checkpoint-5000/
+checkpoint-20000/
+eval_results.json
+word_embeddings_layernorm.pt
+word_embeddings.pt
+train_results.json
+checkpoint-25000/
+all_results.json
+checkpoint-10000/

config.json

@@ -0,0 +1,39 @@
+{
+  "_name_or_path": "bigscience/bloom-560m",
+  "adapters": {
+    "adapters": {},
+    "config_map": {},
+    "fusion_config_map": {},
+    "fusions": {}
+  },
+  "apply_residual_connection_post_layernorm": false,
+  "architectures": [
+    "BloomForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "attention_softmax_in_fp32": true,
+  "bias_dropout_fusion": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_dropout": 0.0,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "masked_softmax_fusion": true,
+  "model_type": "bloom",
+  "n_head": 16,
+  "n_inner": null,
+  "n_layer": 24,
+  "offset_alibi": 100,
+  "pad_token_id": 3,
+  "pretraining_tp": 1,
+  "seq_length": 2048,
+  "skip_bias_add": true,
+  "skip_bias_add_qkv": false,
+  "slow_but_exact": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.20.0.dev0",
+  "unk_token_id": 0,
+  "use_cache": true,
+  "vocab_size": 250880
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a47f3efc641d356f99e7da022403c8ec5c69b65f550c74aca7cd6aaa6e77b7e1
+size 2236955191

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>"}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f6efc66e73f1fd69da4f436e48befb519fdff3fe18910850c1d41bd862293a5
+size 14500443

tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "<pad>", "add_prefix_space": false, "name_or_path": "bigscience/bloom-560m", "special_tokens_map_file": null, "padding_side": "left", "tokenizer_class": "BloomTokenizer"}

trainer_state.json

@@ -0,0 +1,365 @@
+{
+  "best_metric": 1.6942951679229736,
+  "best_model_checkpoint": "/home/lily/nos6/bloom-adapt/exps_ru/bigscience/bloom-560m_ru_adpt_continual-pretrain_original-frozen_100_000samples/checkpoint-25000",
+  "epoch": 1.462073805485701,
+  "global_step": 25000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.8e-05,
+      "loss": 2.6927,
+      "step": 500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.6e-05,
+      "loss": 2.3553,
+      "step": 1000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.4e-05,
+      "loss": 2.2227,
+      "step": 1500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.200000000000001e-05,
+      "loss": 2.1588,
+      "step": 2000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9e-05,
+      "loss": 2.1143,
+      "step": 2500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 8.800000000000001e-05,
+      "loss": 2.0865,
+      "step": 3000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 8.6e-05,
+      "loss": 2.034,
+      "step": 3500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 8.4e-05,
+      "loss": 2.0258,
+      "step": 4000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 8.2e-05,
+      "loss": 2.001,
+      "step": 4500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 8e-05,
+      "loss": 1.9695,
+      "step": 5000
+    },
+    {
+      "epoch": 0.29,
+      "eval_loss": 1.9710078239440918,
+      "eval_runtime": 488.0694,
+      "eval_samples_per_second": 13.633,
+      "eval_steps_per_second": 3.409,
+      "step": 5000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 7.800000000000001e-05,
+      "loss": 1.9591,
+      "step": 5500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 7.6e-05,
+      "loss": 1.9426,
+      "step": 6000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 7.4e-05,
+      "loss": 1.9172,
+      "step": 6500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 7.2e-05,
+      "loss": 1.9401,
+      "step": 7000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 7e-05,
+      "loss": 1.8924,
+      "step": 7500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 6.800000000000001e-05,
+      "loss": 1.8851,
+      "step": 8000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 6.6e-05,
+      "loss": 1.8811,
+      "step": 8500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 6.400000000000001e-05,
+      "loss": 1.8569,
+      "step": 9000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 6.2e-05,
+      "loss": 1.8477,
+      "step": 9500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 6e-05,
+      "loss": 1.854,
+      "step": 10000
+    },
+    {
+      "epoch": 0.58,
+      "eval_loss": 1.8423771858215332,
+      "eval_runtime": 487.9063,
+      "eval_samples_per_second": 13.638,
+      "eval_steps_per_second": 3.41,
+      "step": 10000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 5.8e-05,
+      "loss": 1.8315,
+      "step": 10500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.6000000000000006e-05,
+      "loss": 1.825,
+      "step": 11000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.4000000000000005e-05,
+      "loss": 1.8114,
+      "step": 11500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 5.2000000000000004e-05,
+      "loss": 1.8065,
+      "step": 12000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5e-05,
+      "loss": 1.8002,
+      "step": 12500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.8e-05,
+      "loss": 1.7874,
+      "step": 13000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.600000000000001e-05,
+      "loss": 1.778,
+      "step": 13500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.4000000000000006e-05,
+      "loss": 1.7802,
+      "step": 14000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.2e-05,
+      "loss": 1.7596,
+      "step": 14500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4e-05,
+      "loss": 1.7594,
+      "step": 15000
+    },
+    {
+      "epoch": 0.88,
+      "eval_loss": 1.7678033113479614,
+      "eval_runtime": 487.9075,
+      "eval_samples_per_second": 13.638,
+      "eval_steps_per_second": 3.41,
+      "step": 15000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.8e-05,
+      "loss": 1.7445,
+      "step": 15500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.6e-05,
+      "loss": 1.7464,
+      "step": 16000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.4000000000000007e-05,
+      "loss": 1.7426,
+      "step": 16500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.2000000000000005e-05,
+      "loss": 1.7352,
+      "step": 17000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3e-05,
+      "loss": 1.6449,
+      "step": 17500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 2.8000000000000003e-05,
+      "loss": 1.6259,
+      "step": 18000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 2.6000000000000002e-05,
+      "loss": 1.6189,
+      "step": 18500
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 2.4e-05,
+      "loss": 1.6132,
+      "step": 19000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 2.2000000000000003e-05,
+      "loss": 1.6124,
+      "step": 19500
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 2e-05,
+      "loss": 1.6194,
+      "step": 20000
+    },
+    {
+      "epoch": 1.17,
+      "eval_loss": 1.722885251045227,
+      "eval_runtime": 487.7818,
+      "eval_samples_per_second": 13.641,
+      "eval_steps_per_second": 3.411,
+      "step": 20000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.8e-05,
+      "loss": 1.5974,
+      "step": 20500
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.6000000000000003e-05,
+      "loss": 1.6006,
+      "step": 21000
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.4000000000000001e-05,
+      "loss": 1.591,
+      "step": 21500
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.2e-05,
+      "loss": 1.5985,
+      "step": 22000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1e-05,
+      "loss": 1.5898,
+      "step": 22500
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 1.5857,
+      "step": 23000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 6e-06,
+      "loss": 1.5802,
+      "step": 23500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 1.5797,
+      "step": 24000
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 1.5817,
+      "step": 24500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.0,
+      "loss": 1.5837,
+      "step": 25000
+    },
+    {
+      "epoch": 1.46,
+      "eval_loss": 1.6942951679229736,
+      "eval_runtime": 488.0731,
+      "eval_samples_per_second": 13.633,
+      "eval_steps_per_second": 3.409,
+      "step": 25000
+    },
+    {
+      "epoch": 1.46,
+      "step": 25000,
+      "total_flos": 3.7147907956565606e+17,
+      "train_loss": 1.8233512182617186,
+      "train_runtime": 45032.5538,
+      "train_samples_per_second": 4.441,
+      "train_steps_per_second": 0.555
+    }
+  ],
+  "max_steps": 25000,
+  "num_train_epochs": 2,
+  "total_flos": 3.7147907956565606e+17,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b871752f3c441f14587f9f603bdfd761fdee1a838b97feeccda6721e152361a5
+size 3375