Upload folder using huggingface_hub

106_128_e3_3e-5/adapter_config.json

@@ -0,0 +1,39 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "ibm-granite/granite-3.3-8b-base",
+  "bias": "none",
+  "corda_config": null,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 256,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 128,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "up_proj",
+    "v_proj",
+    "q_proj",
+    "gate_proj",
+    "o_proj",
+    "k_proj",
+    "down_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

106_128_e3_3e-5/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fef1f733c9e844c9c8e1afb3503527aff904ed2c2d84bc1558a8accad30ec156
+size 791751704

106_128_e3_3e-5/latest

@@ -0,0 +1 @@
+global_step774

106_128_e3_3e-5/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a52ee6c59a64cd16d7f0fe4c04314afdcf48673e44c28d4647194d10afaf5fec
+size 16389

106_128_e3_3e-5/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca1c448ff0a7023102ff0716ab421a54d294d85acc5f4ba563c6e26a28fc89a4
+size 16389

106_128_e3_3e-5/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e619bf91a82d3efc74b79b28550f919309eed7ce09716c20367dbb88a448f6dc
+size 16389

106_128_e3_3e-5/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a935e1736c4c9a91c383dbc7796f712fb15d903bfdb1de27722d0e4af0b7fd65
+size 16389

106_128_e3_3e-5/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fadf8f910ce6ae97e28d1a65581843f397354e248e3256309cdaed77351e4a71
+size 16389

106_128_e3_3e-5/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69dbf0849f994d4f3906bedf2c6cbc61a66a31b2cc708376e1f33f7f2a8a3d4b
+size 16389

106_128_e3_3e-5/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b0d9e769807873b2527e280b285aba7482c28977d8cd98e01f35862ffd0f3ac
+size 16389

106_128_e3_3e-5/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ebf6810152ac8e6ae7918161f94876d3f3e821ebfd7262436fee967b0f55c72e
+size 16389

106_128_e3_3e-5/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a35ea21a712c7d513d50b5a81296ea7f712219ae38a046e7526e98537fed71d4
+size 1401

106_128_e3_3e-5/special_tokens_map.json

@@ -0,0 +1,45 @@
+{
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "<fim_prefix>",
+    "<fim_middle>",
+    "<fim_suffix>",
+    "<fim_pad>",
+    "<filename>",
+    "<gh_stars>",
+    "<issue_start>",
+    "<issue_comment>",
+    "<issue_closed>",
+    "<jupyter_start>",
+    "<jupyter_text>",
+    "<jupyter_code>",
+    "<jupyter_output>",
+    "<empty_output>",
+    "<commit_before>",
+    "<commit_msg>",
+    "<commit_after>",
+    "<reponame>"
+  ],
+  "bos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<reponame>",
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

106_128_e3_3e-5/tokenizer_config.json

@@ -0,0 +1,188 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<fim_prefix>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "<fim_middle>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<fim_suffix>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "4": {
+      "content": "<fim_pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "5": {
+      "content": "<filename>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "6": {
+      "content": "<gh_stars>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "7": {
+      "content": "<issue_start>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "8": {
+      "content": "<issue_comment>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "9": {
+      "content": "<issue_closed>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "10": {
+      "content": "<jupyter_start>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "11": {
+      "content": "<jupyter_text>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "12": {
+      "content": "<jupyter_code>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "13": {
+      "content": "<jupyter_output>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "14": {
+      "content": "<empty_output>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "15": {
+      "content": "<commit_before>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "16": {
+      "content": "<commit_msg>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "17": {
+      "content": "<commit_after>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "18": {
+      "content": "<reponame>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "<fim_prefix>",
+    "<fim_middle>",
+    "<fim_suffix>",
+    "<fim_pad>",
+    "<filename>",
+    "<gh_stars>",
+    "<issue_start>",
+    "<issue_comment>",
+    "<issue_closed>",
+    "<jupyter_start>",
+    "<jupyter_text>",
+    "<jupyter_code>",
+    "<jupyter_output>",
+    "<empty_output>",
+    "<commit_before>",
+    "<commit_msg>",
+    "<commit_after>",
+    "<reponame>"
+  ],
+  "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|endoftext|>",
+  "extra_special_tokens": {},
+  "model_max_length": 8192,
+  "pad_token": "<reponame>",
+  "padding_side": "left",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>",
+  "vocab_size": 49152
+}

106_128_e3_3e-5/trainer_state.json

@@ -0,0 +1,1112 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 774,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01937984496124031,
+      "grad_norm": 0.9739851951599121,
+      "learning_rate": 3.076923076923077e-06,
+      "loss": 1.3898,
+      "step": 5
+    },
+    {
+      "epoch": 0.03875968992248062,
+      "grad_norm": 0.9063084125518799,
+      "learning_rate": 6.923076923076923e-06,
+      "loss": 1.4055,
+      "step": 10
+    },
+    {
+      "epoch": 0.05813953488372093,
+      "grad_norm": 0.6522149443626404,
+      "learning_rate": 1.076923076923077e-05,
+      "loss": 1.3956,
+      "step": 15
+    },
+    {
+      "epoch": 0.07751937984496124,
+      "grad_norm": 0.5418137311935425,
+      "learning_rate": 1.4615384615384615e-05,
+      "loss": 1.4017,
+      "step": 20
+    },
+    {
+      "epoch": 0.09689922480620156,
+      "grad_norm": 1.1983472108840942,
+      "learning_rate": 1.8461538461538465e-05,
+      "loss": 1.367,
+      "step": 25
+    },
+    {
+      "epoch": 0.11627906976744186,
+      "grad_norm": 0.5370506644248962,
+      "learning_rate": 2.230769230769231e-05,
+      "loss": 1.3501,
+      "step": 30
+    },
+    {
+      "epoch": 0.13565891472868216,
+      "grad_norm": 0.671127200126648,
+      "learning_rate": 2.6153846153846157e-05,
+      "loss": 1.1992,
+      "step": 35
+    },
+    {
+      "epoch": 0.15503875968992248,
+      "grad_norm": 0.5312590599060059,
+      "learning_rate": 3e-05,
+      "loss": 1.2472,
+      "step": 40
+    },
+    {
+      "epoch": 0.1744186046511628,
+      "grad_norm": 0.5088062882423401,
+      "learning_rate": 2.999657461170462e-05,
+      "loss": 1.176,
+      "step": 45
+    },
+    {
+      "epoch": 0.1937984496124031,
+      "grad_norm": 0.5682260394096375,
+      "learning_rate": 2.9986300011256473e-05,
+      "loss": 1.1622,
+      "step": 50
+    },
+    {
+      "epoch": 0.2131782945736434,
+      "grad_norm": 0.5732645988464355,
+      "learning_rate": 2.9969180891255046e-05,
+      "loss": 1.1143,
+      "step": 55
+    },
+    {
+      "epoch": 0.23255813953488372,
+      "grad_norm": 0.5810739398002625,
+      "learning_rate": 2.9945225070318105e-05,
+      "loss": 1.079,
+      "step": 60
+    },
+    {
+      "epoch": 0.25193798449612403,
+      "grad_norm": 0.6778976321220398,
+      "learning_rate": 2.9914443489510807e-05,
+      "loss": 1.0716,
+      "step": 65
+    },
+    {
+      "epoch": 0.2713178294573643,
+      "grad_norm": 0.6784929633140564,
+      "learning_rate": 2.987685020734869e-05,
+      "loss": 1.0723,
+      "step": 70
+    },
+    {
+      "epoch": 0.29069767441860467,
+      "grad_norm": 0.6986741423606873,
+      "learning_rate": 2.9832462393376926e-05,
+      "loss": 1.0114,
+      "step": 75
+    },
+    {
+      "epoch": 0.31007751937984496,
+      "grad_norm": 0.9799211025238037,
+      "learning_rate": 2.978130032032864e-05,
+      "loss": 1.016,
+      "step": 80
+    },
+    {
+      "epoch": 0.32945736434108525,
+      "grad_norm": 0.9505375623703003,
+      "learning_rate": 2.972338735486598e-05,
+      "loss": 0.9864,
+      "step": 85
+    },
+    {
+      "epoch": 0.3488372093023256,
+      "grad_norm": 0.8457907438278198,
+      "learning_rate": 2.965874994690816e-05,
+      "loss": 0.9082,
+      "step": 90
+    },
+    {
+      "epoch": 0.3682170542635659,
+      "grad_norm": 0.9450764656066895,
+      "learning_rate": 2.9587417617551268e-05,
+      "loss": 0.912,
+      "step": 95
+    },
+    {
+      "epoch": 0.3875968992248062,
+      "grad_norm": 1.0455824136734009,
+      "learning_rate": 2.9509422945585443e-05,
+      "loss": 0.8568,
+      "step": 100
+    },
+    {
+      "epoch": 0.4069767441860465,
+      "grad_norm": 0.8769596219062805,
+      "learning_rate": 2.9424801552615546e-05,
+      "loss": 0.8403,
+      "step": 105
+    },
+    {
+      "epoch": 0.4263565891472868,
+      "grad_norm": 1.096492886543274,
+      "learning_rate": 2.9333592086792113e-05,
+      "loss": 0.7782,
+      "step": 110
+    },
+    {
+      "epoch": 0.44573643410852715,
+      "grad_norm": 1.0260639190673828,
+      "learning_rate": 2.923583620516003e-05,
+      "loss": 0.8136,
+      "step": 115
+    },
+    {
+      "epoch": 0.46511627906976744,
+      "grad_norm": 0.9539602994918823,
+      "learning_rate": 2.9131578554633e-05,
+      "loss": 0.7806,
+      "step": 120
+    },
+    {
+      "epoch": 0.4844961240310077,
+      "grad_norm": 1.0650639533996582,
+      "learning_rate": 2.9020866751602467e-05,
+      "loss": 0.7141,
+      "step": 125
+    },
+    {
+      "epoch": 0.5038759689922481,
+      "grad_norm": 1.1741199493408203,
+      "learning_rate": 2.8903751360190327e-05,
+      "loss": 0.7295,
+      "step": 130
+    },
+    {
+      "epoch": 0.5232558139534884,
+      "grad_norm": 1.0028592348098755,
+      "learning_rate": 2.8780285869155376e-05,
+      "loss": 0.7298,
+      "step": 135
+    },
+    {
+      "epoch": 0.5426356589147286,
+      "grad_norm": 1.060347080230713,
+      "learning_rate": 2.8650526667464e-05,
+      "loss": 0.6636,
+      "step": 140
+    },
+    {
+      "epoch": 0.562015503875969,
+      "grad_norm": 1.1886528730392456,
+      "learning_rate": 2.8514533018536286e-05,
+      "loss": 0.6226,
+      "step": 145
+    },
+    {
+      "epoch": 0.5813953488372093,
+      "grad_norm": 1.1285793781280518,
+      "learning_rate": 2.8372367033179344e-05,
+      "loss": 0.6439,
+      "step": 150
+    },
+    {
+      "epoch": 0.6007751937984496,
+      "grad_norm": 1.253207802772522,
+      "learning_rate": 2.8224093641220134e-05,
+      "loss": 0.6008,
+      "step": 155
+    },
+    {
+      "epoch": 0.6201550387596899,
+      "grad_norm": 1.133418321609497,
+      "learning_rate": 2.806978056185084e-05,
+      "loss": 0.595,
+      "step": 160
+    },
+    {
+      "epoch": 0.6395348837209303,
+      "grad_norm": 1.1310876607894897,
+      "learning_rate": 2.7909498272700244e-05,
+      "loss": 0.6002,
+      "step": 165
+    },
+    {
+      "epoch": 0.6589147286821705,
+      "grad_norm": 1.3232929706573486,
+      "learning_rate": 2.774331997764531e-05,
+      "loss": 0.5768,
+      "step": 170
+    },
+    {
+      "epoch": 0.6782945736434108,
+      "grad_norm": 1.1505982875823975,
+      "learning_rate": 2.7571321573377607e-05,
+      "loss": 0.5796,
+      "step": 175
+    },
+    {
+      "epoch": 0.6976744186046512,
+      "grad_norm": 1.3877726793289185,
+      "learning_rate": 2.7393581614739924e-05,
+      "loss": 0.6397,
+      "step": 180
+    },
+    {
+      "epoch": 0.7170542635658915,
+      "grad_norm": 1.7206389904022217,
+      "learning_rate": 2.7210181278848773e-05,
+      "loss": 0.5304,
+      "step": 185
+    },
+    {
+      "epoch": 0.7364341085271318,
+      "grad_norm": 1.2046453952789307,
+      "learning_rate": 2.702120432801935e-05,
+      "loss": 0.49,
+      "step": 190
+    },
+    {
+      "epoch": 0.7558139534883721,
+      "grad_norm": 1.2877296209335327,
+      "learning_rate": 2.6826737071509716e-05,
+      "loss": 0.5068,
+      "step": 195
+    },
+    {
+      "epoch": 0.7751937984496124,
+      "grad_norm": 1.383624792098999,
+      "learning_rate": 2.662686832610178e-05,
+      "loss": 0.4724,
+      "step": 200
+    },
+    {
+      "epoch": 0.7945736434108527,
+      "grad_norm": 1.4844017028808594,
+      "learning_rate": 2.6421689375537015e-05,
+      "loss": 0.4677,
+      "step": 205
+    },
+    {
+      "epoch": 0.813953488372093,
+      "grad_norm": 1.3249423503875732,
+      "learning_rate": 2.621129392882553e-05,
+      "loss": 0.4579,
+      "step": 210
+    },
+    {
+      "epoch": 0.8333333333333334,
+      "grad_norm": 1.277148723602295,
+      "learning_rate": 2.5995778077447393e-05,
+      "loss": 0.4796,
+      "step": 215
+    },
+    {
+      "epoch": 0.8527131782945736,
+      "grad_norm": 1.4020274877548218,
+      "learning_rate": 2.5775240251465914e-05,
+      "loss": 0.4176,
+      "step": 220
+    },
+    {
+      "epoch": 0.872093023255814,
+      "grad_norm": 1.3385308980941772,
+      "learning_rate": 2.554978117457279e-05,
+      "loss": 0.4502,
+      "step": 225
+    },
+    {
+      "epoch": 0.8914728682170543,
+      "grad_norm": 1.2949692010879517,
+      "learning_rate": 2.531950381808578e-05,
+      "loss": 0.4106,
+      "step": 230
+    },
+    {
+      "epoch": 0.9108527131782945,
+      "grad_norm": 1.3922860622406006,
+      "learning_rate": 2.5084513353919753e-05,
+      "loss": 0.4261,
+      "step": 235
+    },
+    {
+      "epoch": 0.9302325581395349,
+      "grad_norm": 1.4007540941238403,
+      "learning_rate": 2.4844917106552774e-05,
+      "loss": 0.4134,
+      "step": 240
+    },
+    {
+      "epoch": 0.9496124031007752,
+      "grad_norm": 1.2588896751403809,
+      "learning_rate": 2.4600824504009028e-05,
+      "loss": 0.3909,
+      "step": 245
+    },
+    {
+      "epoch": 0.9689922480620154,
+      "grad_norm": 1.764664888381958,
+      "learning_rate": 2.4352347027881003e-05,
+      "loss": 0.3932,
+      "step": 250
+    },
+    {
+      "epoch": 0.9883720930232558,
+      "grad_norm": 1.5046838521957397,
+      "learning_rate": 2.409959816241383e-05,
+      "loss": 0.3743,
+      "step": 255
+    },
+    {
+      "epoch": 1.0077519379844961,
+      "grad_norm": 1.3902626037597656,
+      "learning_rate": 2.38426933426749e-05,
+      "loss": 0.3347,
+      "step": 260
+    },
+    {
+      "epoch": 1.0271317829457365,
+      "grad_norm": 1.211899995803833,
+      "learning_rate": 2.3581749901832546e-05,
+      "loss": 0.3126,
+      "step": 265
+    },
+    {
+      "epoch": 1.0465116279069768,
+      "grad_norm": 1.2012094259262085,
+      "learning_rate": 2.3316887017567844e-05,
+      "loss": 0.3068,
+      "step": 270
+    },
+    {
+      "epoch": 1.0658914728682172,
+      "grad_norm": 1.6160045862197876,
+      "learning_rate": 2.3048225657643945e-05,
+      "loss": 0.2965,
+      "step": 275
+    },
+    {
+      "epoch": 1.0852713178294573,
+      "grad_norm": 1.2792149782180786,
+      "learning_rate": 2.277588852465788e-05,
+      "loss": 0.2838,
+      "step": 280
+    },
+    {
+      "epoch": 1.1046511627906976,
+      "grad_norm": 1.2743208408355713,
+      "learning_rate": 2.25e-05,
+      "loss": 0.2413,
+      "step": 285
+    },
+    {
+      "epoch": 1.124031007751938,
+      "grad_norm": 1.321061372756958,
+      "learning_rate": 2.2220686087046747e-05,
+      "loss": 0.2729,
+      "step": 290
+    },
+    {
+      "epoch": 1.1434108527131783,
+      "grad_norm": 1.399215579032898,
+      "learning_rate": 2.1938074353612533e-05,
+      "loss": 0.2411,
+      "step": 295
+    },
+    {
+      "epoch": 1.1627906976744187,
+      "grad_norm": 1.1990870237350464,
+      "learning_rate": 2.16522938736872e-05,
+      "loss": 0.267,
+      "step": 300
+    },
+    {
+      "epoch": 1.1821705426356588,
+      "grad_norm": 1.3323127031326294,
+      "learning_rate": 2.1363475168485558e-05,
+      "loss": 0.2728,
+      "step": 305
+    },
+    {
+      "epoch": 1.2015503875968991,
+      "grad_norm": 1.3782747983932495,
+      "learning_rate": 2.1071750146835906e-05,
+      "loss": 0.248,
+      "step": 310
+    },
+    {
+      "epoch": 1.2209302325581395,
+      "grad_norm": 1.313757061958313,
+      "learning_rate": 2.077725204493486e-05,
+      "loss": 0.2518,
+      "step": 315
+    },
+    {
+      "epoch": 1.2403100775193798,
+      "grad_norm": 1.3502507209777832,
+      "learning_rate": 2.0480115365495928e-05,
+      "loss": 0.259,
+      "step": 320
+    },
+    {
+      "epoch": 1.2596899224806202,
+      "grad_norm": 1.3070353269577026,
+      "learning_rate": 2.0180475816319613e-05,
+      "loss": 0.2448,
+      "step": 325
+    },
+    {
+      "epoch": 1.2790697674418605,
+      "grad_norm": 1.2860273122787476,
+      "learning_rate": 1.9878470248313208e-05,
+      "loss": 0.2278,
+      "step": 330
+    },
+    {
+      "epoch": 1.2984496124031009,
+      "grad_norm": 1.508518934249878,
+      "learning_rate": 1.957423659298841e-05,
+      "loss": 0.229,
+      "step": 335
+    },
+    {
+      "epoch": 1.3178294573643412,
+      "grad_norm": 1.3453519344329834,
+      "learning_rate": 1.9267913799465486e-05,
+      "loss": 0.2706,
+      "step": 340
+    },
+    {
+      "epoch": 1.3372093023255813,
+      "grad_norm": 1.2780779600143433,
+      "learning_rate": 1.8959641771012648e-05,
+      "loss": 0.2525,
+      "step": 345
+    },
+    {
+      "epoch": 1.3565891472868217,
+      "grad_norm": 1.2714142799377441,
+      "learning_rate": 1.864956130114963e-05,
+      "loss": 0.2535,
+      "step": 350
+    },
+    {
+      "epoch": 1.375968992248062,
+      "grad_norm": 1.146409034729004,
+      "learning_rate": 1.8337814009344716e-05,
+      "loss": 0.1814,
+      "step": 355
+    },
+    {
+      "epoch": 1.3953488372093024,
+      "grad_norm": 1.1344863176345825,
+      "learning_rate": 1.80245422763345e-05,
+      "loss": 0.1901,
+      "step": 360
+    },
+    {
+      "epoch": 1.4147286821705427,
+      "grad_norm": 1.4356662034988403,
+      "learning_rate": 1.7709889179095984e-05,
+      "loss": 0.2277,
+      "step": 365
+    },
+    {
+      "epoch": 1.4341085271317828,
+      "grad_norm": 1.5576506853103638,
+      "learning_rate": 1.739399842550069e-05,
+      "loss": 0.2132,
+      "step": 370
+    },
+    {
+      "epoch": 1.4534883720930232,
+      "grad_norm": 1.0315269231796265,
+      "learning_rate": 1.707701428868061e-05,
+      "loss": 0.2016,
+      "step": 375
+    },
+    {
+      "epoch": 1.4728682170542635,
+      "grad_norm": 1.3095598220825195,
+      "learning_rate": 1.675908154113603e-05,
+      "loss": 0.1903,
+      "step": 380
+    },
+    {
+      "epoch": 1.4922480620155039,
+      "grad_norm": 1.2156273126602173,
+      "learning_rate": 1.6440345388615225e-05,
+      "loss": 0.1841,
+      "step": 385
+    },
+    {
+      "epoch": 1.5116279069767442,
+      "grad_norm": 1.2628873586654663,
+      "learning_rate": 1.6120951403796367e-05,
+      "loss": 0.1762,
+      "step": 390
+    },
+    {
+      "epoch": 1.5310077519379846,
+      "grad_norm": 1.1997421979904175,
+      "learning_rate": 1.5801045459801735e-05,
+      "loss": 0.1755,
+      "step": 395
+    },
+    {
+      "epoch": 1.550387596899225,
+      "grad_norm": 1.106736660003662,
+      "learning_rate": 1.5480773663574832e-05,
+      "loss": 0.1968,
+      "step": 400
+    },
+    {
+      "epoch": 1.5697674418604652,
+      "grad_norm": 1.188226342201233,
+      "learning_rate": 1.51602822891506e-05,
+      "loss": 0.1644,
+      "step": 405
+    },
+    {
+      "epoch": 1.5891472868217056,
+      "grad_norm": 1.037963628768921,
+      "learning_rate": 1.4839717710849403e-05,
+      "loss": 0.1522,
+      "step": 410
+    },
+    {
+      "epoch": 1.6085271317829457,
+      "grad_norm": 1.5000929832458496,
+      "learning_rate": 1.4519226336425172e-05,
+      "loss": 0.1663,
+      "step": 415
+    },
+    {
+      "epoch": 1.627906976744186,
+      "grad_norm": 1.1199333667755127,
+      "learning_rate": 1.4198954540198266e-05,
+      "loss": 0.1603,
+      "step": 420
+    },
+    {
+      "epoch": 1.6472868217054264,
+      "grad_norm": 1.512349247932434,
+      "learning_rate": 1.3879048596203637e-05,
+      "loss": 0.1816,
+      "step": 425
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 1.1847162246704102,
+      "learning_rate": 1.3559654611384774e-05,
+      "loss": 0.1474,
+      "step": 430
+    },
+    {
+      "epoch": 1.6860465116279069,
+      "grad_norm": 1.1964776515960693,
+      "learning_rate": 1.3240918458863975e-05,
+      "loss": 0.1427,
+      "step": 435
+    },
+    {
+      "epoch": 1.7054263565891472,
+      "grad_norm": 1.2912464141845703,
+      "learning_rate": 1.2922985711319387e-05,
+      "loss": 0.1266,
+      "step": 440
+    },
+    {
+      "epoch": 1.7248062015503876,
+      "grad_norm": 1.2497581243515015,
+      "learning_rate": 1.2606001574499316e-05,
+      "loss": 0.1545,
+      "step": 445
+    },
+    {
+      "epoch": 1.744186046511628,
+      "grad_norm": 1.2297734022140503,
+      "learning_rate": 1.2290110820904019e-05,
+      "loss": 0.1547,
+      "step": 450
+    },
+    {
+      "epoch": 1.7635658914728682,
+      "grad_norm": 1.1310908794403076,
+      "learning_rate": 1.1975457723665505e-05,
+      "loss": 0.1347,
+      "step": 455
+    },
+    {
+      "epoch": 1.7829457364341086,
+      "grad_norm": 1.133622407913208,
+      "learning_rate": 1.1662185990655285e-05,
+      "loss": 0.1302,
+      "step": 460
+    },
+    {
+      "epoch": 1.802325581395349,
+      "grad_norm": 0.9459003210067749,
+      "learning_rate": 1.135043869885037e-05,
+      "loss": 0.1302,
+      "step": 465
+    },
+    {
+      "epoch": 1.8217054263565893,
+      "grad_norm": 1.0880463123321533,
+      "learning_rate": 1.1040358228987353e-05,
+      "loss": 0.1272,
+      "step": 470
+    },
+    {
+      "epoch": 1.8410852713178296,
+      "grad_norm": 0.9948514103889465,
+      "learning_rate": 1.0732086200534515e-05,
+      "loss": 0.1233,
+      "step": 475
+    },
+    {
+      "epoch": 1.8604651162790697,
+      "grad_norm": 1.2051372528076172,
+      "learning_rate": 1.0425763407011594e-05,
+      "loss": 0.1314,
+      "step": 480
+    },
+    {
+      "epoch": 1.87984496124031,
+      "grad_norm": 1.1070996522903442,
+      "learning_rate": 1.0121529751686796e-05,
+      "loss": 0.1245,
+      "step": 485
+    },
+    {
+      "epoch": 1.8992248062015504,
+      "grad_norm": 1.384994387626648,
+      "learning_rate": 9.819524183680388e-06,
+      "loss": 0.1286,
+      "step": 490
+    },
+    {
+      "epoch": 1.9186046511627906,
+      "grad_norm": 1.1956759691238403,
+      "learning_rate": 9.519884634504074e-06,
+      "loss": 0.1014,
+      "step": 495
+    },
+    {
+      "epoch": 1.937984496124031,
+      "grad_norm": 0.9163130521774292,
+      "learning_rate": 9.222747955065138e-06,
+      "loss": 0.1212,
+      "step": 500
+    },
+    {
+      "epoch": 1.9573643410852712,
+      "grad_norm": 1.0861150026321411,
+      "learning_rate": 8.928249853164093e-06,
+      "loss": 0.1278,
+      "step": 505
+    },
+    {
+      "epoch": 1.9767441860465116,
+      "grad_norm": 1.0453848838806152,
+      "learning_rate": 8.636524831514436e-06,
+      "loss": 0.1127,
+      "step": 510
+    },
+    {
+      "epoch": 1.996124031007752,
+      "grad_norm": 1.0486317873001099,
+      "learning_rate": 8.347706126312798e-06,
+      "loss": 0.1068,
+      "step": 515
+    },
+    {
+      "epoch": 2.0155038759689923,
+      "grad_norm": 0.9646169543266296,
+      "learning_rate": 8.061925646387474e-06,
+      "loss": 0.0758,
+      "step": 520
+    },
+    {
+      "epoch": 2.0348837209302326,
+      "grad_norm": 1.2811315059661865,
+      "learning_rate": 7.779313912953262e-06,
+      "loss": 0.0875,
+      "step": 525
+    },
+    {
+      "epoch": 2.054263565891473,
+      "grad_norm": 0.9795780181884766,
+      "learning_rate": 7.500000000000004e-06,
+      "loss": 0.0992,
+      "step": 530
+    },
+    {
+      "epoch": 2.0736434108527133,
+      "grad_norm": 0.9730613231658936,
+      "learning_rate": 7.2241114753421245e-06,
+      "loss": 0.0797,
+      "step": 535
+    },
+    {
+      "epoch": 2.0930232558139537,
+      "grad_norm": 0.9051758646965027,
+      "learning_rate": 6.951774342356058e-06,
+      "loss": 0.0892,
+      "step": 540
+    },
+    {
+      "epoch": 2.112403100775194,
+      "grad_norm": 0.9356362223625183,
+      "learning_rate": 6.6831129824321565e-06,
+      "loss": 0.0844,
+      "step": 545
+    },
+    {
+      "epoch": 2.1317829457364343,
+      "grad_norm": 0.9345132112503052,
+      "learning_rate": 6.4182500981674585e-06,
+      "loss": 0.0829,
+      "step": 550
+    },
+    {
+      "epoch": 2.1511627906976742,
+      "grad_norm": 0.79043048620224,
+      "learning_rate": 6.157306657325105e-06,
+      "loss": 0.0796,
+      "step": 555
+    },
+    {
+      "epoch": 2.1705426356589146,
+      "grad_norm": 1.0362004041671753,
+      "learning_rate": 5.900401837586169e-06,
+      "loss": 0.084,
+      "step": 560
+    },
+    {
+      "epoch": 2.189922480620155,
+      "grad_norm": 0.7386631369590759,
+      "learning_rate": 5.647652972118998e-06,
+      "loss": 0.0727,
+      "step": 565
+    },
+    {
+      "epoch": 2.2093023255813953,
+      "grad_norm": 0.7612237334251404,
+      "learning_rate": 5.399175495990975e-06,
+      "loss": 0.0774,
+      "step": 570
+    },
+    {
+      "epoch": 2.2286821705426356,
+      "grad_norm": 0.9880594611167908,
+      "learning_rate": 5.155082893447228e-06,
+      "loss": 0.0716,
+      "step": 575
+    },
+    {
+      "epoch": 2.248062015503876,
+      "grad_norm": 1.3928707838058472,
+      "learning_rate": 4.915486646080247e-06,
+      "loss": 0.0797,
+      "step": 580
+    },
+    {
+      "epoch": 2.2674418604651163,
+      "grad_norm": 0.9496927261352539,
+      "learning_rate": 4.680496181914222e-06,
+      "loss": 0.0752,
+      "step": 585
+    },
+    {
+      "epoch": 2.2868217054263567,
+      "grad_norm": 0.8892703652381897,
+      "learning_rate": 4.450218825427205e-06,
+      "loss": 0.0768,
+      "step": 590
+    },
+    {
+      "epoch": 2.306201550387597,
+      "grad_norm": 0.7403905391693115,
+      "learning_rate": 4.224759748534088e-06,
+      "loss": 0.074,
+      "step": 595
+    },
+    {
+      "epoch": 2.3255813953488373,
+      "grad_norm": 0.8606444001197815,
+      "learning_rate": 4.004221922552608e-06,
+      "loss": 0.0749,
+      "step": 600
+    },
+    {
+      "epoch": 2.3449612403100777,
+      "grad_norm": 0.7143356204032898,
+      "learning_rate": 3.7887060711744718e-06,
+      "loss": 0.0722,
+      "step": 605
+    },
+    {
+      "epoch": 2.3643410852713176,
+      "grad_norm": 0.8410319685935974,
+      "learning_rate": 3.5783106244629837e-06,
+      "loss": 0.0608,
+      "step": 610
+    },
+    {
+      "epoch": 2.383720930232558,
+      "grad_norm": 0.8149115443229675,
+      "learning_rate": 3.373131673898225e-06,
+      "loss": 0.0699,
+      "step": 615
+    },
+    {
+      "epoch": 2.4031007751937983,
+      "grad_norm": 0.874718427658081,
+      "learning_rate": 3.173262928490284e-06,
+      "loss": 0.0733,
+      "step": 620
+    },
+    {
+      "epoch": 2.4224806201550386,
+      "grad_norm": 0.8637734651565552,
+      "learning_rate": 2.9787956719806526e-06,
+      "loss": 0.0815,
+      "step": 625
+    },
+    {
+      "epoch": 2.441860465116279,
+      "grad_norm": 0.8122116923332214,
+      "learning_rate": 2.789818721151227e-06,
+      "loss": 0.0671,
+      "step": 630
+    },
+    {
+      "epoch": 2.4612403100775193,
+      "grad_norm": 0.9417327642440796,
+      "learning_rate": 2.60641838526008e-06,
+      "loss": 0.0631,
+      "step": 635
+    },
+    {
+      "epoch": 2.4806201550387597,
+      "grad_norm": 0.8427896499633789,
+      "learning_rate": 2.428678426622391e-06,
+      "loss": 0.07,
+      "step": 640
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 2.2559187412261963,
+      "learning_rate": 2.2566800223546913e-06,
+      "loss": 0.0678,
+      "step": 645
+    },
+    {
+      "epoch": 2.5193798449612403,
+      "grad_norm": 0.7065724730491638,
+      "learning_rate": 2.0905017272997562e-06,
+      "loss": 0.0682,
+      "step": 650
+    },
+    {
+      "epoch": 2.5387596899224807,
+      "grad_norm": 0.7164368629455566,
+      "learning_rate": 1.930219438149159e-06,
+      "loss": 0.0683,
+      "step": 655
+    },
+    {
+      "epoch": 2.558139534883721,
+      "grad_norm": 0.6012015342712402,
+      "learning_rate": 1.7759063587798647e-06,
+      "loss": 0.0681,
+      "step": 660
+    },
+    {
+      "epoch": 2.5775193798449614,
+      "grad_norm": 0.8413594365119934,
+      "learning_rate": 1.627632966820658e-06,
+      "loss": 0.0632,
+      "step": 665
+    },
+    {
+      "epoch": 2.5968992248062017,
+      "grad_norm": 1.0626535415649414,
+      "learning_rate": 1.4854669814637145e-06,
+      "loss": 0.0699,
+      "step": 670
+    },
+    {
+      "epoch": 2.616279069767442,
+      "grad_norm": 0.6342124342918396,
+      "learning_rate": 1.3494733325360027e-06,
+      "loss": 0.0629,
+      "step": 675
+    },
+    {
+      "epoch": 2.6356589147286824,
+      "grad_norm": 0.7284303307533264,
+      "learning_rate": 1.2197141308446248e-06,
+      "loss": 0.0676,
+      "step": 680
+    },
+    {
+      "epoch": 2.6550387596899228,
+      "grad_norm": 0.7012219429016113,
+      "learning_rate": 1.0962486398096755e-06,
+      "loss": 0.0604,
+      "step": 685
+    },
+    {
+      "epoch": 2.6744186046511627,
+      "grad_norm": 0.7915197014808655,
+      "learning_rate": 9.791332483975346e-07,
+      "loss": 0.0623,
+      "step": 690
+    },
+    {
+      "epoch": 2.693798449612403,
+      "grad_norm": 0.774778425693512,
+      "learning_rate": 8.684214453669991e-07,
+      "loss": 0.0622,
+      "step": 695
+    },
+    {
+      "epoch": 2.7131782945736433,
+      "grad_norm": 0.6751202940940857,
+      "learning_rate": 7.641637948399705e-07,
+      "loss": 0.0607,
+      "step": 700
+    },
+    {
+      "epoch": 2.7325581395348837,
+      "grad_norm": 0.7893778085708618,
+      "learning_rate": 6.664079132078881e-07,
+      "loss": 0.0635,
+      "step": 705
+    },
+    {
+      "epoch": 2.751937984496124,
+      "grad_norm": 0.7164852023124695,
+      "learning_rate": 5.751984473844568e-07,
+      "loss": 0.0592,
+      "step": 710
+    },
+    {
+      "epoch": 2.7713178294573644,
+      "grad_norm": 0.6754224896430969,
+      "learning_rate": 4.905770544145583e-07,
+      "loss": 0.0617,
+      "step": 715
+    },
+    {
+      "epoch": 2.7906976744186047,
+      "grad_norm": 0.7164446711540222,
+      "learning_rate": 4.1258238244873404e-07,
+      "loss": 0.0616,
+      "step": 720
+    },
+    {
+      "epoch": 2.810077519379845,
+      "grad_norm": 0.6280542612075806,
+      "learning_rate": 3.4125005309183997e-07,
+      "loss": 0.0613,
+      "step": 725
+    },
+    {
+      "epoch": 2.8294573643410854,
+      "grad_norm": 0.7230589985847473,
+      "learning_rate": 2.766126451340184e-07,
+      "loss": 0.0654,
+      "step": 730
+    },
+    {
+      "epoch": 2.8488372093023253,
+      "grad_norm": 0.6544074416160583,
+      "learning_rate": 2.1869967967136196e-07,
+      "loss": 0.0663,
+      "step": 735
+    },
+    {
+      "epoch": 2.8682170542635657,
+      "grad_norm": 0.7592392563819885,
+      "learning_rate": 1.6753760662307217e-07,
+      "loss": 0.0653,
+      "step": 740
+    },
+    {
+      "epoch": 2.887596899224806,
+      "grad_norm": 0.7099024653434753,
+      "learning_rate": 1.231497926513092e-07,
+      "loss": 0.0594,
+      "step": 745
+    },
+    {
+      "epoch": 2.9069767441860463,
+      "grad_norm": 0.6078741550445557,
+      "learning_rate": 8.5556510489197e-08,
+      "loss": 0.0597,
+      "step": 750
+    },
+    {
+      "epoch": 2.9263565891472867,
+      "grad_norm": 0.7022770047187805,
+      "learning_rate": 5.4774929681896125e-08,
+      "loss": 0.0591,
+      "step": 755
+    },
+    {
+      "epoch": 2.945736434108527,
+      "grad_norm": 0.7501327395439148,
+      "learning_rate": 3.081910874495497e-08,
+      "loss": 0.0708,
+      "step": 760
+    },
+    {
+      "epoch": 2.9651162790697674,
+      "grad_norm": 0.7156089544296265,
+      "learning_rate": 1.3699988743525738e-08,
+      "loss": 0.0657,
+      "step": 765
+    },
+    {
+      "epoch": 2.9844961240310077,
+      "grad_norm": 0.8015995025634766,
+      "learning_rate": 3.4253882953816374e-09,
+      "loss": 0.0647,
+      "step": 770
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 774,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 2000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 9.772156245120123e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

106_128_e3_3e-5/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72216c22c8a6e9d8f3ff74f476407fc102bc15ea4f9089dbca32ef86e0b41028
+size 8209

106_128_e3_3e-5/zero_to_fp32.py

@@ -0,0 +1,604 @@
+#!/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 1, 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())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        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_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    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, exclude_frozen_parameters):
+    """
+    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,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+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 _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+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() if _has_callable(shape, 'numel') else math.prod(shape)
+            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,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _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:
+        if pair[1] in state_dict:
+            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,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _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:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    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``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    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, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    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``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    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("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    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,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)