Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ 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
+pytorch_model.bin filter=lfs diff=lfs merge=lfs -text

config.json

@@ -0,0 +1,287 @@
+{
+  "_commit_hash": "fa9c7486149855fffd7f495aed428f2b025f4fd5",
+  "_name_or_path": "Salesforce/instructblip-flan-t5-xl",
+  "architectures": [
+    "InstructBlipForConditionalGeneration"
+  ],
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "is_encoder_decoder": true,
+  "model_type": "instructblip",
+  "num_query_tokens": 32,
+  "qformer_config": {
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": null,
+    "attention_probs_dropout_prob": 0.1,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "classifier_dropout": null,
+    "cross_attention_frequency": 2,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "early_stopping": false,
+    "encoder_hidden_size": 1408,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": null,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_act": "gelu",
+    "hidden_dropout_prob": 0.1,
+    "hidden_size": 768,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "intermediate_size": 3072,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-12,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "max_position_embeddings": 512,
+    "min_length": 0,
+    "model_type": "instructblip_qformer",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 12,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_hidden_layers": 12,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": 0,
+    "position_embedding_type": "absolute",
+    "prefix": null,
+    "problem_type": null,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": null,
+    "torchscript": false,
+    "transformers_version": "4.29.2",
+    "typical_p": 1.0,
+    "use_bfloat16": false,
+    "vocab_size": 30623
+  },
+  "text_config": {
+    "_from_model_config": false,
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": [
+      "T5ForConditionalGeneration"
+    ],
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": 1,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "d_ff": 5120,
+    "d_kv": 64,
+    "d_model": 2048,
+    "decoder_start_token_id": 0,
+    "dense_act_fn": "gelu",
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "dropout_rate": 0.1,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": 1,
+    "exponential_decay_length_penalty": null,
+    "feed_forward_proj": "gated-gelu",
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_factor": 1.0,
+    "is_decoder": false,
+    "is_encoder_decoder": true,
+    "is_gated_act": true,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_epsilon": 1e-06,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "t5",
+    "n_positions": 512,
+    "no_repeat_ngram_size": 0,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_decoder_layers": 24,
+    "num_heads": 32,
+    "num_layers": 24,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_past": true,
+    "output_scores": false,
+    "pad_token_id": 0,
+    "prefix": null,
+    "problem_type": null,
+    "pruned_heads": {},
+    "relative_attention_max_distance": 128,
+    "relative_attention_num_buckets": 32,
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": {
+      "summarization": {
+        "early_stopping": true,
+        "length_penalty": 2.0,
+        "max_length": 200,
+        "min_length": 30,
+        "no_repeat_ngram_size": 3,
+        "num_beams": 4,
+        "prefix": "summarize: "
+      },
+      "translation_en_to_de": {
+        "early_stopping": true,
+        "max_length": 300,
+        "num_beams": 4,
+        "prefix": "translate English to German: "
+      },
+      "translation_en_to_fr": {
+        "early_stopping": true,
+        "max_length": 300,
+        "num_beams": 4,
+        "prefix": "translate English to French: "
+      },
+      "translation_en_to_ro": {
+        "early_stopping": true,
+        "max_length": 300,
+        "num_beams": 4,
+        "prefix": "translate English to Romanian: "
+      }
+    },
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": false,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": "float32",
+    "torchscript": false,
+    "transformers_version": "4.29.2",
+    "typical_p": 1.0,
+    "use_bfloat16": false,
+    "use_cache": true,
+    "vocab_size": 32128
+  },
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": null,
+  "use_decoder_only_language_model": false,
+  "vision_config": {
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": null,
+    "attention_dropout": 0.0,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "dropout": 0.0,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": null,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_act": "gelu",
+    "hidden_size": 1408,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "image_size": 224,
+    "initializer_factor": 1.0,
+    "initializer_range": 1e-10,
+    "intermediate_size": 6144,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-05,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "instructblip_vision_model",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 16,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_channels": 3,
+    "num_hidden_layers": 39,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": null,
+    "patch_size": 14,
+    "prefix": null,
+    "problem_type": null,
+    "projection_dim": 512,
+    "pruned_heads": {},
+    "qkv_bias": true,
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": null,
+    "torchscript": false,
+    "transformers_version": "4.29.2",
+    "typical_p": 1.0,
+    "use_bfloat16": false
+  }
+}

latest

@@ -0,0 +1 @@
+global_step1500

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f6d48d314bcb8a19ada769dc9af8fcfcfc9be42d221b50f4f55235f7fd1896b
+size 8046567875

rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c68a25d37b0b194d5f39d5814c524d4177edcc93c7db7a776c49cac6f44b2613
+size 19539

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:53b883a036b069efa1e59a3729009e30533185b026ee03a684d4d765323f4314
+size 19539

rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b1376b9b767573d999b4d567a1c07f8456006c2b071a5b322801e3ed02dfdf8
+size 19539

rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f539f195bfec0ddcb112938aaa247d81e7cb19dd57d0680d12b71ed2261fde0d
+size 19539

rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d11c9946f779d146f0ff48b5d0ea3ed57eb631a5ef63ee95eff09fe85a0965ed
+size 19539

rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a94e70f86a5e40d54d6d12a1a47f30d0dbc1c94f37160060be4979a2da03cef
+size 19539

trainer_state.json

@@ -0,0 +1,874 @@
+{
+  "best_metric": 0.11633333333333333,
+  "best_model_checkpoint": "checkpoints/instruct_BLIP2_deepSpeed_t5xl_unfreeze_Qformer_Projection_Encoder_DecoderLLM_QV_weight_no_instructqformer/checkpoint-1200",
+  "epoch": 1.271994912020352,
+  "global_step": 1500,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.04,
+      "eval_accuracy": 0.10066666666666667,
+      "eval_avg_bleuScore": 0.028052001625299455,
+      "eval_loss": 1.821718692779541,
+      "eval_rouge1_fmeasure": 0.39566364884376526,
+      "eval_rouge1_precision": 0.40851742029190063,
+      "eval_rouge1_recall": 0.40402013063430786,
+      "eval_rouge2_fmeasure": 0.11902458965778351,
+      "eval_rouge2_precision": 0.12462873011827469,
+      "eval_rouge2_recall": 0.12082352489233017,
+      "eval_rougeL_fmeasure": 0.3664987087249756,
+      "eval_rougeL_precision": 0.37793973088264465,
+      "eval_rougeL_recall": 0.374777227640152,
+      "eval_rougeLsum_fmeasure": 0.36648938059806824,
+      "eval_rougeLsum_precision": 0.3779275715351105,
+      "eval_rougeLsum_recall": 0.3747769892215729,
+      "eval_runtime": 497.9069,
+      "eval_samples_per_second": 6.025,
+      "eval_steps_per_second": 0.084,
+      "step": 50
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.10066666666666667,
+      "epoch": 0.04,
+      "step": 50
+    },
+    {
+      "epoch": 0.08,
+      "eval_accuracy": 0.10766666666666666,
+      "eval_avg_bleuScore": 0.027600931445757548,
+      "eval_loss": 1.7757344245910645,
+      "eval_rouge1_fmeasure": 0.4066351652145386,
+      "eval_rouge1_precision": 0.42435961961746216,
+      "eval_rouge1_recall": 0.4121783375740051,
+      "eval_rouge2_fmeasure": 0.1213102638721466,
+      "eval_rouge2_precision": 0.12911927700042725,
+      "eval_rouge2_recall": 0.12066558748483658,
+      "eval_rougeL_fmeasure": 0.375918984413147,
+      "eval_rougeL_precision": 0.3917166292667389,
+      "eval_rougeL_recall": 0.3817567229270935,
+      "eval_rougeLsum_fmeasure": 0.37593305110931396,
+      "eval_rougeLsum_precision": 0.3917284309864044,
+      "eval_rougeLsum_recall": 0.38177403807640076,
+      "eval_runtime": 665.5797,
+      "eval_samples_per_second": 4.507,
+      "eval_steps_per_second": 0.063,
+      "step": 100
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.10766666666666666,
+      "epoch": 0.08,
+      "step": 100
+    },
+    {
+      "epoch": 0.13,
+      "eval_accuracy": 0.10366666666666667,
+      "eval_avg_bleuScore": 0.027236416570842265,
+      "eval_loss": 1.761132836341858,
+      "eval_rouge1_fmeasure": 0.4095948040485382,
+      "eval_rouge1_precision": 0.4314381182193756,
+      "eval_rouge1_recall": 0.4098112881183624,
+      "eval_rouge2_fmeasure": 0.12273810803890228,
+      "eval_rouge2_precision": 0.13252894580364227,
+      "eval_rouge2_recall": 0.12072920054197311,
+      "eval_rougeL_fmeasure": 0.3781542181968689,
+      "eval_rougeL_precision": 0.3977195620536804,
+      "eval_rougeL_recall": 0.37909412384033203,
+      "eval_rougeLsum_fmeasure": 0.3781542181968689,
+      "eval_rougeLsum_precision": 0.3977195620536804,
+      "eval_rougeLsum_recall": 0.37909412384033203,
+      "eval_runtime": 682.2715,
+      "eval_samples_per_second": 4.397,
+      "eval_steps_per_second": 0.062,
+      "step": 150
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.10766666666666666,
+      "epoch": 0.13,
+      "step": 150
+    },
+    {
+      "epoch": 0.17,
+      "eval_accuracy": 0.111,
+      "eval_avg_bleuScore": 0.02802771261582772,
+      "eval_loss": 1.752191424369812,
+      "eval_rouge1_fmeasure": 0.4127008020877838,
+      "eval_rouge1_precision": 0.43622446060180664,
+      "eval_rouge1_recall": 0.4123985171318054,
+      "eval_rouge2_fmeasure": 0.12286161631345749,
+      "eval_rouge2_precision": 0.13331134617328644,
+      "eval_rouge2_recall": 0.11954713612794876,
+      "eval_rougeL_fmeasure": 0.38183602690696716,
+      "eval_rougeL_precision": 0.4026814103126526,
+      "eval_rougeL_recall": 0.38258570432662964,
+      "eval_rougeLsum_fmeasure": 0.38183602690696716,
+      "eval_rougeLsum_precision": 0.4026814103126526,
+      "eval_rougeLsum_recall": 0.38258570432662964,
+      "eval_runtime": 651.8851,
+      "eval_samples_per_second": 4.602,
+      "eval_steps_per_second": 0.064,
+      "step": 200
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.111,
+      "epoch": 0.17,
+      "step": 200
+    },
+    {
+      "epoch": 0.21,
+      "eval_accuracy": 0.105,
+      "eval_avg_bleuScore": 0.027756770220274726,
+      "eval_loss": 1.7442187070846558,
+      "eval_rouge1_fmeasure": 0.4091792702674866,
+      "eval_rouge1_precision": 0.4254671335220337,
+      "eval_rouge1_recall": 0.4135805368423462,
+      "eval_rouge2_fmeasure": 0.12273108959197998,
+      "eval_rouge2_precision": 0.13018882274627686,
+      "eval_rouge2_recall": 0.1215042695403099,
+      "eval_rougeL_fmeasure": 0.3773505389690399,
+      "eval_rougeL_precision": 0.3915799856185913,
+      "eval_rougeL_recall": 0.38237836956977844,
+      "eval_rougeLsum_fmeasure": 0.3773505389690399,
+      "eval_rougeLsum_precision": 0.3915799856185913,
+      "eval_rougeLsum_recall": 0.38237836956977844,
+      "eval_runtime": 537.4239,
+      "eval_samples_per_second": 5.582,
+      "eval_steps_per_second": 0.078,
+      "step": 250
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.111,
+      "epoch": 0.21,
+      "step": 250
+    },
+    {
+      "epoch": 0.25,
+      "eval_accuracy": 0.10933333333333334,
+      "eval_avg_bleuScore": 0.029238753189643224,
+      "eval_loss": 1.7391133308410645,
+      "eval_rouge1_fmeasure": 0.4145018756389618,
+      "eval_rouge1_precision": 0.4305371046066284,
+      "eval_rouge1_recall": 0.418173223733902,
+      "eval_rouge2_fmeasure": 0.12484551221132278,
+      "eval_rouge2_precision": 0.13209865987300873,
+      "eval_rouge2_recall": 0.123726025223732,
+      "eval_rougeL_fmeasure": 0.3840600848197937,
+      "eval_rougeL_precision": 0.39804303646087646,
+      "eval_rougeL_recall": 0.388366162776947,
+      "eval_rougeLsum_fmeasure": 0.38407862186431885,
+      "eval_rougeLsum_precision": 0.3980652689933777,
+      "eval_rougeLsum_recall": 0.38838204741477966,
+      "eval_runtime": 636.8392,
+      "eval_samples_per_second": 4.711,
+      "eval_steps_per_second": 0.066,
+      "step": 300
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.111,
+      "epoch": 0.25,
+      "step": 300
+    },
+    {
+      "epoch": 0.3,
+      "eval_accuracy": 0.10833333333333334,
+      "eval_avg_bleuScore": 0.02794275293002526,
+      "eval_loss": 1.7383359670639038,
+      "eval_rouge1_fmeasure": 0.4148719012737274,
+      "eval_rouge1_precision": 0.4366380274295807,
+      "eval_rouge1_recall": 0.41247957944869995,
+      "eval_rouge2_fmeasure": 0.12334803491830826,
+      "eval_rouge2_precision": 0.13329564034938812,
+      "eval_rouge2_recall": 0.11994606256484985,
+      "eval_rougeL_fmeasure": 0.3842999339103699,
+      "eval_rougeL_precision": 0.403416246175766,
+      "eval_rougeL_recall": 0.3831061124801636,
+      "eval_rougeLsum_fmeasure": 0.38431844115257263,
+      "eval_rougeLsum_precision": 0.4034384787082672,
+      "eval_rougeLsum_recall": 0.3831219971179962,
+      "eval_runtime": 613.4876,
+      "eval_samples_per_second": 4.89,
+      "eval_steps_per_second": 0.068,
+      "step": 350
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.111,
+      "epoch": 0.3,
+      "step": 350
+    },
+    {
+      "epoch": 0.34,
+      "eval_accuracy": 0.11166666666666666,
+      "eval_avg_bleuScore": 0.027422180160880088,
+      "eval_loss": 1.7323437929153442,
+      "eval_rouge1_fmeasure": 0.41112449765205383,
+      "eval_rouge1_precision": 0.42366519570350647,
+      "eval_rouge1_recall": 0.42089518904685974,
+      "eval_rouge2_fmeasure": 0.12152101844549179,
+      "eval_rouge2_precision": 0.1277269721031189,
+      "eval_rouge2_recall": 0.1217166930437088,
+      "eval_rougeL_fmeasure": 0.38071784377098083,
+      "eval_rougeL_precision": 0.39161595702171326,
+      "eval_rougeL_recall": 0.3908142149448395,
+      "eval_rougeLsum_fmeasure": 0.380736380815506,
+      "eval_rougeLsum_precision": 0.3916381895542145,
+      "eval_rougeLsum_recall": 0.39083006978034973,
+      "eval_runtime": 832.3704,
+      "eval_samples_per_second": 3.604,
+      "eval_steps_per_second": 0.05,
+      "step": 400
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.11166666666666666,
+      "epoch": 0.34,
+      "step": 400
+    },
+    {
+      "epoch": 0.38,
+      "eval_accuracy": 0.113,
+      "eval_avg_bleuScore": 0.02941073496143023,
+      "eval_loss": 1.7295702695846558,
+      "eval_rouge1_fmeasure": 0.4182838201522827,
+      "eval_rouge1_precision": 0.4307253062725067,
+      "eval_rouge1_recall": 0.42410746216773987,
+      "eval_rouge2_fmeasure": 0.12393321841955185,
+      "eval_rouge2_precision": 0.1300639659166336,
+      "eval_rouge2_recall": 0.1235329881310463,
+      "eval_rougeL_fmeasure": 0.38701331615448,
+      "eval_rougeL_precision": 0.3976545035839081,
+      "eval_rougeL_recall": 0.3932779133319855,
+      "eval_rougeLsum_fmeasure": 0.38703182339668274,
+      "eval_rougeLsum_precision": 0.3976767063140869,
+      "eval_rougeLsum_recall": 0.3932937681674957,
+      "eval_runtime": 644.9911,
+      "eval_samples_per_second": 4.651,
+      "eval_steps_per_second": 0.065,
+      "step": 450
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.113,
+      "epoch": 0.38,
+      "step": 450
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.07225348510492e-05,
+      "loss": 1.4583,
+      "step": 500
+    },
+    {
+      "epoch": 0.42,
+      "eval_accuracy": 0.112,
+      "eval_avg_bleuScore": 0.028446144399543603,
+      "eval_loss": 1.7309530973434448,
+      "eval_rouge1_fmeasure": 0.41481727361679077,
+      "eval_rouge1_precision": 0.4275760352611542,
+      "eval_rouge1_recall": 0.4214574098587036,
+      "eval_rouge2_fmeasure": 0.12195532023906708,
+      "eval_rouge2_precision": 0.1282820999622345,
+      "eval_rouge2_recall": 0.12207218259572983,
+      "eval_rougeL_fmeasure": 0.38423991203308105,
+      "eval_rougeL_precision": 0.39524608850479126,
+      "eval_rougeL_recall": 0.39129284024238586,
+      "eval_rougeLsum_fmeasure": 0.38425952196121216,
+      "eval_rougeLsum_precision": 0.3952717185020447,
+      "eval_rougeLsum_recall": 0.3913087248802185,
+      "eval_runtime": 753.8166,
+      "eval_samples_per_second": 3.98,
+      "eval_steps_per_second": 0.056,
+      "step": 500
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.113,
+      "epoch": 0.42,
+      "step": 500
+    },
+    {
+      "epoch": 0.47,
+      "eval_accuracy": 0.109,
+      "eval_avg_bleuScore": 0.02862334846953551,
+      "eval_loss": 1.7262461185455322,
+      "eval_rouge1_fmeasure": 0.41011935472488403,
+      "eval_rouge1_precision": 0.4244203269481659,
+      "eval_rouge1_recall": 0.4158960282802582,
+      "eval_rouge2_fmeasure": 0.12124613672494888,
+      "eval_rouge2_precision": 0.12845604121685028,
+      "eval_rouge2_recall": 0.1203385517001152,
+      "eval_rougeL_fmeasure": 0.37967798113822937,
+      "eval_rougeL_precision": 0.3920755982398987,
+      "eval_rougeL_recall": 0.3858604431152344,
+      "eval_rougeLsum_fmeasure": 0.3796946704387665,
+      "eval_rougeLsum_precision": 0.39209315180778503,
+      "eval_rougeLsum_recall": 0.38587629795074463,
+      "eval_runtime": 521.071,
+      "eval_samples_per_second": 5.757,
+      "eval_steps_per_second": 0.081,
+      "step": 550
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.113,
+      "epoch": 0.47,
+      "step": 550
+    },
+    {
+      "epoch": 0.51,
+      "eval_accuracy": 0.11133333333333334,
+      "eval_avg_bleuScore": 0.028583366366724174,
+      "eval_loss": 1.7252265214920044,
+      "eval_rouge1_fmeasure": 0.4135696589946747,
+      "eval_rouge1_precision": 0.42282983660697937,
+      "eval_rouge1_recall": 0.42099812626838684,
+      "eval_rouge2_fmeasure": 0.12228861451148987,
+      "eval_rouge2_precision": 0.1273748129606247,
+      "eval_rouge2_recall": 0.12293525040149689,
+      "eval_rougeL_fmeasure": 0.3833465278148651,
+      "eval_rougeL_precision": 0.391093373298645,
+      "eval_rougeL_recall": 0.39074525237083435,
+      "eval_rougeLsum_fmeasure": 0.38336315751075745,
+      "eval_rougeLsum_precision": 0.39111092686653137,
+      "eval_rougeLsum_recall": 0.3907611072063446,
+      "eval_runtime": 444.5441,
+      "eval_samples_per_second": 6.748,
+      "eval_steps_per_second": 0.094,
+      "step": 600
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.113,
+      "epoch": 0.51,
+      "step": 600
+    },
+    {
+      "epoch": 0.55,
+      "eval_accuracy": 0.11166666666666666,
+      "eval_avg_bleuScore": 0.0276417871316274,
+      "eval_loss": 1.7287226915359497,
+      "eval_rouge1_fmeasure": 0.41999372839927673,
+      "eval_rouge1_precision": 0.4347474277019501,
+      "eval_rouge1_recall": 0.41890957951545715,
+      "eval_rouge2_fmeasure": 0.1226092204451561,
+      "eval_rouge2_precision": 0.12977544963359833,
+      "eval_rouge2_recall": 0.12063863128423691,
+      "eval_rougeL_fmeasure": 0.38852831721305847,
+      "eval_rougeL_precision": 0.401227742433548,
+      "eval_rougeL_recall": 0.3881692588329315,
+      "eval_rougeLsum_fmeasure": 0.3885473608970642,
+      "eval_rougeLsum_precision": 0.40125155448913574,
+      "eval_rougeLsum_recall": 0.38818514347076416,
+      "eval_runtime": 416.9212,
+      "eval_samples_per_second": 7.196,
+      "eval_steps_per_second": 0.101,
+      "step": 650
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.113,
+      "epoch": 0.55,
+      "step": 650
+    },
+    {
+      "epoch": 0.59,
+      "eval_accuracy": 0.11166666666666666,
+      "eval_avg_bleuScore": 0.028842177291711173,
+      "eval_loss": 1.7268632650375366,
+      "eval_rouge1_fmeasure": 0.4184536039829254,
+      "eval_rouge1_precision": 0.4310172498226166,
+      "eval_rouge1_recall": 0.4187770485877991,
+      "eval_rouge2_fmeasure": 0.12251175940036774,
+      "eval_rouge2_precision": 0.12893937528133392,
+      "eval_rouge2_recall": 0.12104275822639465,
+      "eval_rougeL_fmeasure": 0.38783329725265503,
+      "eval_rougeL_precision": 0.398539274930954,
+      "eval_rougeL_recall": 0.3887461721897125,
+      "eval_rougeLsum_fmeasure": 0.38785290718078613,
+      "eval_rougeLsum_precision": 0.3985649049282074,
+      "eval_rougeLsum_recall": 0.38876205682754517,
+      "eval_runtime": 417.6826,
+      "eval_samples_per_second": 7.182,
+      "eval_steps_per_second": 0.101,
+      "step": 700
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.113,
+      "epoch": 0.59,
+      "step": 700
+    },
+    {
+      "epoch": 0.64,
+      "eval_accuracy": 0.115,
+      "eval_avg_bleuScore": 0.0290608666613698,
+      "eval_loss": 1.721261739730835,
+      "eval_rouge1_fmeasure": 0.4215930998325348,
+      "eval_rouge1_precision": 0.43335336446762085,
+      "eval_rouge1_recall": 0.4254607856273651,
+      "eval_rouge2_fmeasure": 0.12333113700151443,
+      "eval_rouge2_precision": 0.1291794627904892,
+      "eval_rouge2_recall": 0.123162180185318,
+      "eval_rougeL_fmeasure": 0.39064252376556396,
+      "eval_rougeL_precision": 0.40073809027671814,
+      "eval_rougeL_recall": 0.39462339878082275,
+      "eval_rougeLsum_fmeasure": 0.39064013957977295,
+      "eval_rougeLsum_precision": 0.4007423222064972,
+      "eval_rougeLsum_recall": 0.39460596442222595,
+      "eval_runtime": 448.6827,
+      "eval_samples_per_second": 6.686,
+      "eval_steps_per_second": 0.094,
+      "step": 750
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.115,
+      "epoch": 0.64,
+      "step": 750
+    },
+    {
+      "epoch": 0.68,
+      "eval_accuracy": 0.11366666666666667,
+      "eval_avg_bleuScore": 0.02823698183397452,
+      "eval_loss": 1.719179630279541,
+      "eval_rouge1_fmeasure": 0.4255302846431732,
+      "eval_rouge1_precision": 0.44179973006248474,
+      "eval_rouge1_recall": 0.4235004186630249,
+      "eval_rouge2_fmeasure": 0.12516146898269653,
+      "eval_rouge2_precision": 0.13273605704307556,
+      "eval_rouge2_recall": 0.12307706475257874,
+      "eval_rougeL_fmeasure": 0.3939012885093689,
+      "eval_rougeL_precision": 0.4078863561153412,
+      "eval_rougeL_recall": 0.39276742935180664,
+      "eval_rougeLsum_fmeasure": 0.3939012885093689,
+      "eval_rougeLsum_precision": 0.4078863561153412,
+      "eval_rougeLsum_recall": 0.39276742935180664,
+      "eval_runtime": 428.3483,
+      "eval_samples_per_second": 7.004,
+      "eval_steps_per_second": 0.098,
+      "step": 800
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.115,
+      "epoch": 0.68,
+      "step": 800
+    },
+    {
+      "epoch": 0.72,
+      "eval_accuracy": 0.11433333333333333,
+      "eval_avg_bleuScore": 0.030075389464696248,
+      "eval_loss": 1.716699242591858,
+      "eval_rouge1_fmeasure": 0.42475032806396484,
+      "eval_rouge1_precision": 0.4392147660255432,
+      "eval_rouge1_recall": 0.4252479374408722,
+      "eval_rouge2_fmeasure": 0.12556101381778717,
+      "eval_rouge2_precision": 0.13259626924991608,
+      "eval_rouge2_recall": 0.12433860450983047,
+      "eval_rougeL_fmeasure": 0.39261123538017273,
+      "eval_rougeL_precision": 0.40510380268096924,
+      "eval_rougeL_recall": 0.3936094641685486,
+      "eval_rougeLsum_fmeasure": 0.3926297128200531,
+      "eval_rougeLsum_precision": 0.40512600541114807,
+      "eval_rougeLsum_recall": 0.39362531900405884,
+      "eval_runtime": 416.1761,
+      "eval_samples_per_second": 7.208,
+      "eval_steps_per_second": 0.101,
+      "step": 850
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.115,
+      "epoch": 0.72,
+      "step": 850
+    },
+    {
+      "epoch": 0.76,
+      "eval_accuracy": 0.11333333333333333,
+      "eval_avg_bleuScore": 0.02911172941327095,
+      "eval_loss": 1.7144726514816284,
+      "eval_rouge1_fmeasure": 0.4232616424560547,
+      "eval_rouge1_precision": 0.43622541427612305,
+      "eval_rouge1_recall": 0.4255277216434479,
+      "eval_rouge2_fmeasure": 0.12597259879112244,
+      "eval_rouge2_precision": 0.1326662003993988,
+      "eval_rouge2_recall": 0.12559087574481964,
+      "eval_rougeL_fmeasure": 0.3917687237262726,
+      "eval_rougeL_precision": 0.4028547704219818,
+      "eval_rougeL_recall": 0.39450860023498535,
+      "eval_rougeLsum_fmeasure": 0.39178499579429626,
+      "eval_rougeLsum_precision": 0.40287142992019653,
+      "eval_rougeLsum_recall": 0.3945244550704956,
+      "eval_runtime": 418.7629,
+      "eval_samples_per_second": 7.164,
+      "eval_steps_per_second": 0.1,
+      "step": 900
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.115,
+      "epoch": 0.76,
+      "step": 900
+    },
+    {
+      "epoch": 0.81,
+      "eval_accuracy": 0.11366666666666667,
+      "eval_avg_bleuScore": 0.02877427616963784,
+      "eval_loss": 1.7141094207763672,
+      "eval_rouge1_fmeasure": 0.42343342304229736,
+      "eval_rouge1_precision": 0.4381575882434845,
+      "eval_rouge1_recall": 0.4245162308216095,
+      "eval_rouge2_fmeasure": 0.12420167028903961,
+      "eval_rouge2_precision": 0.13117600977420807,
+      "eval_rouge2_recall": 0.12346965819597244,
+      "eval_rougeL_fmeasure": 0.3917766809463501,
+      "eval_rougeL_precision": 0.4044066071510315,
+      "eval_rougeL_recall": 0.39342838525772095,
+      "eval_rougeLsum_fmeasure": 0.39179524779319763,
+      "eval_rougeLsum_precision": 0.4044288396835327,
+      "eval_rougeLsum_recall": 0.3934442400932312,
+      "eval_runtime": 412.8924,
+      "eval_samples_per_second": 7.266,
+      "eval_steps_per_second": 0.102,
+      "step": 950
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.115,
+      "epoch": 0.81,
+      "step": 950
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.0001,
+      "loss": 1.3251,
+      "step": 1000
+    },
+    {
+      "epoch": 0.85,
+      "eval_accuracy": 0.114,
+      "eval_avg_bleuScore": 0.028698996223509313,
+      "eval_loss": 1.7141406536102295,
+      "eval_rouge1_fmeasure": 0.4231337010860443,
+      "eval_rouge1_precision": 0.43527066707611084,
+      "eval_rouge1_recall": 0.42574357986450195,
+      "eval_rouge2_fmeasure": 0.12320207059383392,
+      "eval_rouge2_precision": 0.12941864132881165,
+      "eval_rouge2_recall": 0.12254573404788971,
+      "eval_rougeL_fmeasure": 0.3913033902645111,
+      "eval_rougeL_precision": 0.40169021487236023,
+      "eval_rougeL_recall": 0.39426353573799133,
+      "eval_rougeLsum_fmeasure": 0.3913209140300751,
+      "eval_rougeLsum_precision": 0.40170982480049133,
+      "eval_rougeLsum_recall": 0.39427945017814636,
+      "eval_runtime": 411.3837,
+      "eval_samples_per_second": 7.292,
+      "eval_steps_per_second": 0.102,
+      "step": 1000
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.115,
+      "epoch": 0.85,
+      "step": 1000
+    },
+    {
+      "epoch": 0.89,
+      "eval_accuracy": 0.115,
+      "eval_avg_bleuScore": 0.028069007659951847,
+      "eval_loss": 1.7145625352859497,
+      "eval_rouge1_fmeasure": 0.4248062074184418,
+      "eval_rouge1_precision": 0.4398679733276367,
+      "eval_rouge1_recall": 0.423473060131073,
+      "eval_rouge2_fmeasure": 0.12515196204185486,
+      "eval_rouge2_precision": 0.13223549723625183,
+      "eval_rouge2_recall": 0.12349213659763336,
+      "eval_rougeL_fmeasure": 0.39318299293518066,
+      "eval_rougeL_precision": 0.405906617641449,
+      "eval_rougeL_recall": 0.39283618330955505,
+      "eval_rougeLsum_fmeasure": 0.39318299293518066,
+      "eval_rougeLsum_precision": 0.405906617641449,
+      "eval_rougeLsum_recall": 0.39283618330955505,
+      "eval_runtime": 423.8761,
+      "eval_samples_per_second": 7.078,
+      "eval_steps_per_second": 0.099,
+      "step": 1050
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.115,
+      "epoch": 0.89,
+      "step": 1050
+    },
+    {
+      "epoch": 0.93,
+      "eval_accuracy": 0.112,
+      "eval_avg_bleuScore": 0.027762405360738436,
+      "eval_loss": 1.7136757373809814,
+      "eval_rouge1_fmeasure": 0.42095834016799927,
+      "eval_rouge1_precision": 0.43534284830093384,
+      "eval_rouge1_recall": 0.42047563195228577,
+      "eval_rouge2_fmeasure": 0.12339958548545837,
+      "eval_rouge2_precision": 0.13038772344589233,
+      "eval_rouge2_recall": 0.12185768783092499,
+      "eval_rougeL_fmeasure": 0.3895176649093628,
+      "eval_rougeL_precision": 0.4018775522708893,
+      "eval_rougeL_recall": 0.3896372616291046,
+      "eval_rougeLsum_fmeasure": 0.3895343542098999,
+      "eval_rougeLsum_precision": 0.40189510583877563,
+      "eval_rougeLsum_recall": 0.38965311646461487,
+      "eval_runtime": 412.9847,
+      "eval_samples_per_second": 7.264,
+      "eval_steps_per_second": 0.102,
+      "step": 1100
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.115,
+      "epoch": 0.93,
+      "step": 1100
+    },
+    {
+      "epoch": 0.98,
+      "eval_accuracy": 0.11233333333333333,
+      "eval_avg_bleuScore": 0.029023348152637483,
+      "eval_loss": 1.7126835584640503,
+      "eval_rouge1_fmeasure": 0.42007914185523987,
+      "eval_rouge1_precision": 0.4302547574043274,
+      "eval_rouge1_recall": 0.42434659600257874,
+      "eval_rouge2_fmeasure": 0.12373975664377213,
+      "eval_rouge2_precision": 0.12897014617919922,
+      "eval_rouge2_recall": 0.1240621954202652,
+      "eval_rougeL_fmeasure": 0.38924336433410645,
+      "eval_rougeL_precision": 0.39777547121047974,
+      "eval_rougeL_recall": 0.39375001192092896,
+      "eval_rougeLsum_fmeasure": 0.389260470867157,
+      "eval_rougeLsum_precision": 0.3977939784526825,
+      "eval_rougeLsum_recall": 0.39376595616340637,
+      "eval_runtime": 416.6217,
+      "eval_samples_per_second": 7.201,
+      "eval_steps_per_second": 0.101,
+      "step": 1150
+    },
+    {
+      "best_epoch": 0,
+      "best_eval_accuracy": 0.115,
+      "epoch": 0.98,
+      "step": 1150
+    },
+    {
+      "epoch": 1.02,
+      "eval_accuracy": 0.11633333333333333,
+      "eval_avg_bleuScore": 0.0286271261125803,
+      "eval_loss": 1.7119219303131104,
+      "eval_rouge1_fmeasure": 0.4223006069660187,
+      "eval_rouge1_precision": 0.43355467915534973,
+      "eval_rouge1_recall": 0.42424923181533813,
+      "eval_rouge2_fmeasure": 0.12183848023414612,
+      "eval_rouge2_precision": 0.12746664881706238,
+      "eval_rouge2_recall": 0.12144064158201218,
+      "eval_rougeL_fmeasure": 0.39173752069473267,
+      "eval_rougeL_precision": 0.4012112319469452,
+      "eval_rougeL_recall": 0.3941992223262787,
+      "eval_rougeLsum_fmeasure": 0.39173752069473267,
+      "eval_rougeLsum_precision": 0.4012112319469452,
+      "eval_rougeLsum_recall": 0.3941992223262787,
+      "eval_runtime": 414.7589,
+      "eval_samples_per_second": 7.233,
+      "eval_steps_per_second": 0.101,
+      "step": 1200
+    },
+    {
+      "best_epoch": 1,
+      "best_eval_accuracy": 0.11633333333333333,
+      "epoch": 1.02,
+      "step": 1200
+    },
+    {
+      "epoch": 1.06,
+      "eval_accuracy": 0.11333333333333333,
+      "eval_avg_bleuScore": 0.027211002017060917,
+      "eval_loss": 1.714453101158142,
+      "eval_rouge1_fmeasure": 0.41994115710258484,
+      "eval_rouge1_precision": 0.43688705563545227,
+      "eval_rouge1_recall": 0.41850513219833374,
+      "eval_rouge2_fmeasure": 0.12223239243030548,
+      "eval_rouge2_precision": 0.13026195764541626,
+      "eval_rouge2_recall": 0.12033700197935104,
+      "eval_rougeL_fmeasure": 0.3890978693962097,
+      "eval_rougeL_precision": 0.4037608504295349,
+      "eval_rougeL_recall": 0.3884038031101227,
+      "eval_rougeLsum_fmeasure": 0.38911449909210205,
+      "eval_rougeLsum_precision": 0.40377840399742126,
+      "eval_rougeLsum_recall": 0.3884196877479553,
+      "eval_runtime": 435.9324,
+      "eval_samples_per_second": 6.882,
+      "eval_steps_per_second": 0.096,
+      "step": 1250
+    },
+    {
+      "best_epoch": 1,
+      "best_eval_accuracy": 0.11633333333333333,
+      "epoch": 1.06,
+      "step": 1250
+    },
+    {
+      "epoch": 1.1,
+      "eval_accuracy": 0.11233333333333333,
+      "eval_avg_bleuScore": 0.02936673008153836,
+      "eval_loss": 1.7116132974624634,
+      "eval_rouge1_fmeasure": 0.4216597378253937,
+      "eval_rouge1_precision": 0.4335962235927582,
+      "eval_rouge1_recall": 0.42250940203666687,
+      "eval_rouge2_fmeasure": 0.12213711440563202,
+      "eval_rouge2_precision": 0.12782621383666992,
+      "eval_rouge2_recall": 0.12156455218791962,
+      "eval_rougeL_fmeasure": 0.3895263671875,
+      "eval_rougeL_precision": 0.39964497089385986,
+      "eval_rougeL_recall": 0.39094799757003784,
+      "eval_rougeLsum_fmeasure": 0.38954439759254456,
+      "eval_rougeLsum_precision": 0.39966580271720886,
+      "eval_rougeLsum_recall": 0.3909638524055481,
+      "eval_runtime": 442.558,
+      "eval_samples_per_second": 6.779,
+      "eval_steps_per_second": 0.095,
+      "step": 1300
+    },
+    {
+      "best_epoch": 1,
+      "best_eval_accuracy": 0.11633333333333333,
+      "epoch": 1.1,
+      "step": 1300
+    },
+    {
+      "epoch": 1.14,
+      "eval_accuracy": 0.11633333333333333,
+      "eval_avg_bleuScore": 0.027918557037909825,
+      "eval_loss": 1.7098476886749268,
+      "eval_rouge1_fmeasure": 0.4242185950279236,
+      "eval_rouge1_precision": 0.4375033378601074,
+      "eval_rouge1_recall": 0.42478516697883606,
+      "eval_rouge2_fmeasure": 0.12248878926038742,
+      "eval_rouge2_precision": 0.12885503470897675,
+      "eval_rouge2_recall": 0.12164236605167389,
+      "eval_rougeL_fmeasure": 0.39410412311553955,
+      "eval_rougeL_precision": 0.40549689531326294,
+      "eval_rougeL_recall": 0.39526131749153137,
+      "eval_rougeLsum_fmeasure": 0.3941212296485901,
+      "eval_rougeLsum_precision": 0.4055154621601105,
+      "eval_rougeLsum_recall": 0.3952771723270416,
+      "eval_runtime": 422.7031,
+      "eval_samples_per_second": 7.097,
+      "eval_steps_per_second": 0.099,
+      "step": 1350
+    },
+    {
+      "best_epoch": 1,
+      "best_eval_accuracy": 0.11633333333333333,
+      "epoch": 1.14,
+      "step": 1350
+    },
+    {
+      "epoch": 1.19,
+      "eval_accuracy": 0.11266666666666666,
+      "eval_avg_bleuScore": 0.027800548608104388,
+      "eval_loss": 1.7062265872955322,
+      "eval_rouge1_fmeasure": 0.42027780413627625,
+      "eval_rouge1_precision": 0.435363233089447,
+      "eval_rouge1_recall": 0.4199471175670624,
+      "eval_rouge2_fmeasure": 0.12210887670516968,
+      "eval_rouge2_precision": 0.12934038043022156,
+      "eval_rouge2_recall": 0.12082366645336151,
+      "eval_rougeL_fmeasure": 0.38862261176109314,
+      "eval_rougeL_precision": 0.40165209770202637,
+      "eval_rougeL_recall": 0.388929158449173,
+      "eval_rougeLsum_fmeasure": 0.3886396884918213,
+      "eval_rougeLsum_precision": 0.4016706645488739,
+      "eval_rougeLsum_recall": 0.3889450132846832,
+      "eval_runtime": 421.8991,
+      "eval_samples_per_second": 7.111,
+      "eval_steps_per_second": 0.1,
+      "step": 1400
+    },
+    {
+      "best_epoch": 1,
+      "best_eval_accuracy": 0.11633333333333333,
+      "epoch": 1.19,
+      "step": 1400
+    },
+    {
+      "epoch": 1.23,
+      "eval_accuracy": 0.113,
+      "eval_avg_bleuScore": 0.026157520641883213,
+      "eval_loss": 1.7059766054153442,
+      "eval_rouge1_fmeasure": 0.4198099374771118,
+      "eval_rouge1_precision": 0.43453213572502136,
+      "eval_rouge1_recall": 0.4178878962993622,
+      "eval_rouge2_fmeasure": 0.1191374883055687,
+      "eval_rouge2_precision": 0.12592871487140656,
+      "eval_rouge2_recall": 0.11757373064756393,
+      "eval_rougeL_fmeasure": 0.3882248103618622,
+      "eval_rougeL_precision": 0.40069010853767395,
+      "eval_rougeL_recall": 0.3872298300266266,
+      "eval_rougeLsum_fmeasure": 0.3882444202899933,
+      "eval_rougeLsum_precision": 0.40071573853492737,
+      "eval_rougeLsum_recall": 0.38724568486213684,
+      "eval_runtime": 425.7259,
+      "eval_samples_per_second": 7.047,
+      "eval_steps_per_second": 0.099,
+      "step": 1450
+    },
+    {
+      "best_epoch": 1,
+      "best_eval_accuracy": 0.11633333333333333,
+      "epoch": 1.23,
+      "step": 1450
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.0001,
+      "loss": 1.2797,
+      "step": 1500
+    },
+    {
+      "epoch": 1.27,
+      "eval_accuracy": 0.114,
+      "eval_avg_bleuScore": 0.026971253817280133,
+      "eval_loss": 1.7088124752044678,
+      "eval_rouge1_fmeasure": 0.42218562960624695,
+      "eval_rouge1_precision": 0.43385228514671326,
+      "eval_rouge1_recall": 0.42346033453941345,
+      "eval_rouge2_fmeasure": 0.12264084815979004,
+      "eval_rouge2_precision": 0.1281544268131256,
+      "eval_rouge2_recall": 0.12249463051557541,
+      "eval_rougeL_fmeasure": 0.3912815749645233,
+      "eval_rougeL_precision": 0.4010494649410248,
+      "eval_rougeL_recall": 0.39323461055755615,
+      "eval_rougeLsum_fmeasure": 0.39129638671875,
+      "eval_rougeLsum_precision": 0.40106332302093506,
+      "eval_rougeLsum_recall": 0.3932504951953888,
+      "eval_runtime": 429.3226,
+      "eval_samples_per_second": 6.988,
+      "eval_steps_per_second": 0.098,
+      "step": 1500
+    },
+    {
+      "best_epoch": 1,
+      "best_eval_accuracy": 0.11633333333333333,
+      "epoch": 1.27,
+      "step": 1500
+    }
+  ],
+  "max_steps": 4716,
+  "num_train_epochs": 4,
+  "total_flos": 2.130248764531802e+22,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fb2690f93601f3894e4e2f7f511834e110a8dbd35cdc3d13e655df1773103c4d
+size 5551

zero_to_fp32.py

@@ -0,0 +1,584 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage == 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # record shared parameters so that they can be recovered based on partners
+        # this is because such parameters holding reference only are not saved by optimizer
+        shared_params = []
+        for param in state_dict["module"]:
+            if param not in [*param_names, *buffer_names]:
+                for share_param in state_dict["module"]:
+                    if (state_dict["module"][share_param].data_ptr() == state_dict["module"][param].data_ptr()
+                            and share_param != param):
+                        shared_params.append([param, share_param])
+                        break
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dicts.append(torch.load(f, map_location=device))
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage == 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage == 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage == 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel()
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, args.output_file)