Upload folder using huggingface_hub

12_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": [
+    "o_proj",
+    "down_proj",
+    "gate_proj",
+    "k_proj",
+    "up_proj",
+    "v_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

12_128_e3_3e-5/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20f03fb0fe7e9a1b8865f66f533257c4bcc6981ceb42a5ea1bf45825b6e4f57c
+size 791751704

12_128_e3_3e-5/latest

@@ -0,0 +1 @@
+global_step744

12_128_e3_3e-5/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6838692ce58749ca5953b6fc78127e4d9623bebe35406a64d6edade3d89edd7
+size 16389

12_128_e3_3e-5/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e52f37ebc450a361d9f701af6f6991ec88a2658c66eca3eeda541c3811d8a6a
+size 16389

12_128_e3_3e-5/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6dcb3556ac9003d54181922cab5be1c5b7568e7174f68baa5945518c8b5a727d
+size 16389

12_128_e3_3e-5/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0213a0541526541aafdedab7c3f4b89b962350182b4ad52eb31edb48d79caaac
+size 16389

12_128_e3_3e-5/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3638a71d9d073c7868c18a9d9316617baa6ccfacd492e19899b4c709f0c9546
+size 16389

12_128_e3_3e-5/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:190fa28f1d5f41a893b103df0de359347a7560c1aea93bab8106139b62937ef5
+size 16389

12_128_e3_3e-5/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d138663b23098803738ec7504089c747510124c8505c9e4221d081633f59f83b
+size 16389

12_128_e3_3e-5/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6d456015afd53cdae706a1b27b638acc1fd598a413111f8a6149ec3cc9b0e8f
+size 16389

12_128_e3_3e-5/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0cc83b5a13903afe6e3a2d9d84bf1e710698d3151306738bc706761ea089d07
+size 1401

12_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
+  }
+}

12_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
+}

12_128_e3_3e-5/trainer_state.json

@@ -0,0 +1,1070 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 744,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.020202020202020204,
+      "grad_norm": 0.9354684352874756,
+      "learning_rate": 3.157894736842105e-06,
+      "loss": 1.2974,
+      "step": 5
+    },
+    {
+      "epoch": 0.04040404040404041,
+      "grad_norm": 0.7772208452224731,
+      "learning_rate": 7.105263157894737e-06,
+      "loss": 1.2806,
+      "step": 10
+    },
+    {
+      "epoch": 0.06060606060606061,
+      "grad_norm": 0.7184535264968872,
+      "learning_rate": 1.1052631578947368e-05,
+      "loss": 1.2889,
+      "step": 15
+    },
+    {
+      "epoch": 0.08080808080808081,
+      "grad_norm": 0.6504032015800476,
+      "learning_rate": 1.5e-05,
+      "loss": 1.256,
+      "step": 20
+    },
+    {
+      "epoch": 0.10101010101010101,
+      "grad_norm": 0.6334185004234314,
+      "learning_rate": 1.894736842105263e-05,
+      "loss": 1.2227,
+      "step": 25
+    },
+    {
+      "epoch": 0.12121212121212122,
+      "grad_norm": 0.6026784777641296,
+      "learning_rate": 2.2894736842105263e-05,
+      "loss": 1.1638,
+      "step": 30
+    },
+    {
+      "epoch": 0.1414141414141414,
+      "grad_norm": 0.5309629440307617,
+      "learning_rate": 2.6842105263157896e-05,
+      "loss": 1.1837,
+      "step": 35
+    },
+    {
+      "epoch": 0.16161616161616163,
+      "grad_norm": 0.6159191727638245,
+      "learning_rate": 2.9999851491644133e-05,
+      "loss": 1.1124,
+      "step": 40
+    },
+    {
+      "epoch": 0.18181818181818182,
+      "grad_norm": 0.6355125904083252,
+      "learning_rate": 2.999465400794852e-05,
+      "loss": 1.1252,
+      "step": 45
+    },
+    {
+      "epoch": 0.20202020202020202,
+      "grad_norm": 0.5908980965614319,
+      "learning_rate": 2.9982034046828747e-05,
+      "loss": 0.9964,
+      "step": 50
+    },
+    {
+      "epoch": 0.2222222222222222,
+      "grad_norm": 0.6541002988815308,
+      "learning_rate": 2.9961997855269662e-05,
+      "loss": 1.0477,
+      "step": 55
+    },
+    {
+      "epoch": 0.24242424242424243,
+      "grad_norm": 0.7609748244285583,
+      "learning_rate": 2.9934555351351374e-05,
+      "loss": 1.0454,
+      "step": 60
+    },
+    {
+      "epoch": 0.26262626262626265,
+      "grad_norm": 0.6675081253051758,
+      "learning_rate": 2.989972011933969e-05,
+      "loss": 0.9803,
+      "step": 65
+    },
+    {
+      "epoch": 0.2828282828282828,
+      "grad_norm": 0.7312514781951904,
+      "learning_rate": 2.9857509402961814e-05,
+      "loss": 0.8957,
+      "step": 70
+    },
+    {
+      "epoch": 0.30303030303030304,
+      "grad_norm": 0.7570792436599731,
+      "learning_rate": 2.9807944096870547e-05,
+      "loss": 0.8469,
+      "step": 75
+    },
+    {
+      "epoch": 0.32323232323232326,
+      "grad_norm": 0.7641143798828125,
+      "learning_rate": 2.9751048736301263e-05,
+      "loss": 0.849,
+      "step": 80
+    },
+    {
+      "epoch": 0.3434343434343434,
+      "grad_norm": 0.9888669848442078,
+      "learning_rate": 2.968685148492677e-05,
+      "loss": 0.8546,
+      "step": 85
+    },
+    {
+      "epoch": 0.36363636363636365,
+      "grad_norm": 0.7715857625007629,
+      "learning_rate": 2.9615384120916058e-05,
+      "loss": 0.797,
+      "step": 90
+    },
+    {
+      "epoch": 0.3838383838383838,
+      "grad_norm": 0.8602557182312012,
+      "learning_rate": 2.9536682021203852e-05,
+      "loss": 0.7839,
+      "step": 95
+    },
+    {
+      "epoch": 0.40404040404040403,
+      "grad_norm": 0.9596008658409119,
+      "learning_rate": 2.945078414397871e-05,
+      "loss": 0.7895,
+      "step": 100
+    },
+    {
+      "epoch": 0.42424242424242425,
+      "grad_norm": 0.9746429324150085,
+      "learning_rate": 2.9357733009398432e-05,
+      "loss": 0.8051,
+      "step": 105
+    },
+    {
+      "epoch": 0.4444444444444444,
+      "grad_norm": 0.8758135437965393,
+      "learning_rate": 2.9257574678542205e-05,
+      "loss": 0.7447,
+      "step": 110
+    },
+    {
+      "epoch": 0.46464646464646464,
+      "grad_norm": 0.911910891532898,
+      "learning_rate": 2.9150358730610013e-05,
+      "loss": 0.6746,
+      "step": 115
+    },
+    {
+      "epoch": 0.48484848484848486,
+      "grad_norm": 0.9135602116584778,
+      "learning_rate": 2.903613823838049e-05,
+      "loss": 0.6394,
+      "step": 120
+    },
+    {
+      "epoch": 0.5050505050505051,
+      "grad_norm": 0.9695019721984863,
+      "learning_rate": 2.8914969741939498e-05,
+      "loss": 0.6439,
+      "step": 125
+    },
+    {
+      "epoch": 0.5252525252525253,
+      "grad_norm": 0.9459332823753357,
+      "learning_rate": 2.8786913220692286e-05,
+      "loss": 0.5921,
+      "step": 130
+    },
+    {
+      "epoch": 0.5454545454545454,
+      "grad_norm": 1.1566177606582642,
+      "learning_rate": 2.86520320636732e-05,
+      "loss": 0.5945,
+      "step": 135
+    },
+    {
+      "epoch": 0.5656565656565656,
+      "grad_norm": 1.1200796365737915,
+      "learning_rate": 2.8510393038167604e-05,
+      "loss": 0.5882,
+      "step": 140
+    },
+    {
+      "epoch": 0.5858585858585859,
+      "grad_norm": 1.0685203075408936,
+      "learning_rate": 2.8362066256661484e-05,
+      "loss": 0.5262,
+      "step": 145
+    },
+    {
+      "epoch": 0.6060606060606061,
+      "grad_norm": 1.014054536819458,
+      "learning_rate": 2.820712514213524e-05,
+      "loss": 0.599,
+      "step": 150
+    },
+    {
+      "epoch": 0.6262626262626263,
+      "grad_norm": 1.2513902187347412,
+      "learning_rate": 2.8045646391718654e-05,
+      "loss": 0.5535,
+      "step": 155
+    },
+    {
+      "epoch": 0.6464646464646465,
+      "grad_norm": 1.1771626472473145,
+      "learning_rate": 2.7877709938725216e-05,
+      "loss": 0.508,
+      "step": 160
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 1.3093689680099487,
+      "learning_rate": 2.770339891308446e-05,
+      "loss": 0.5125,
+      "step": 165
+    },
+    {
+      "epoch": 0.6868686868686869,
+      "grad_norm": 1.2347291707992554,
+      "learning_rate": 2.7522799600191955e-05,
+      "loss": 0.5176,
+      "step": 170
+    },
+    {
+      "epoch": 0.7070707070707071,
+      "grad_norm": 1.10127854347229,
+      "learning_rate": 2.733600139819735e-05,
+      "loss": 0.4945,
+      "step": 175
+    },
+    {
+      "epoch": 0.7272727272727273,
+      "grad_norm": 1.1869375705718994,
+      "learning_rate": 2.7143096773751516e-05,
+      "loss": 0.5005,
+      "step": 180
+    },
+    {
+      "epoch": 0.7474747474747475,
+      "grad_norm": 1.1467169523239136,
+      "learning_rate": 2.6944181216234816e-05,
+      "loss": 0.4636,
+      "step": 185
+    },
+    {
+      "epoch": 0.7676767676767676,
+      "grad_norm": 1.2301661968231201,
+      "learning_rate": 2.673935319048906e-05,
+      "loss": 0.447,
+      "step": 190
+    },
+    {
+      "epoch": 0.7878787878787878,
+      "grad_norm": 1.1257785558700562,
+      "learning_rate": 2.6528714088076624e-05,
+      "loss": 0.3996,
+      "step": 195
+    },
+    {
+      "epoch": 0.8080808080808081,
+      "grad_norm": 1.1837537288665771,
+      "learning_rate": 2.6312368177090742e-05,
+      "loss": 0.4321,
+      "step": 200
+    },
+    {
+      "epoch": 0.8282828282828283,
+      "grad_norm": 1.1720499992370605,
+      "learning_rate": 2.6090422550541983e-05,
+      "loss": 0.3905,
+      "step": 205
+    },
+    {
+      "epoch": 0.8484848484848485,
+      "grad_norm": 1.3978396654129028,
+      "learning_rate": 2.5862987073346308e-05,
+      "loss": 0.3943,
+      "step": 210
+    },
+    {
+      "epoch": 0.8686868686868687,
+      "grad_norm": 1.2642910480499268,
+      "learning_rate": 2.563017432794104e-05,
+      "loss": 0.3404,
+      "step": 215
+    },
+    {
+      "epoch": 0.8888888888888888,
+      "grad_norm": 1.3827550411224365,
+      "learning_rate": 2.5392099558555607e-05,
+      "loss": 0.3728,
+      "step": 220
+    },
+    {
+      "epoch": 0.9090909090909091,
+      "grad_norm": 1.4432976245880127,
+      "learning_rate": 2.5148880614164726e-05,
+      "loss": 0.3618,
+      "step": 225
+    },
+    {
+      "epoch": 0.9292929292929293,
+      "grad_norm": 1.1570686101913452,
+      "learning_rate": 2.4900637890152177e-05,
+      "loss": 0.4401,
+      "step": 230
+    },
+    {
+      "epoch": 0.9494949494949495,
+      "grad_norm": 1.5704818964004517,
+      "learning_rate": 2.464749426871408e-05,
+      "loss": 0.3153,
+      "step": 235
+    },
+    {
+      "epoch": 0.9696969696969697,
+      "grad_norm": 1.3928354978561401,
+      "learning_rate": 2.438957505803123e-05,
+      "loss": 0.3779,
+      "step": 240
+    },
+    {
+      "epoch": 0.98989898989899,
+      "grad_norm": 1.5814493894577026,
+      "learning_rate": 2.4127007930240467e-05,
+      "loss": 0.3543,
+      "step": 245
+    },
+    {
+      "epoch": 1.0080808080808081,
+      "grad_norm": 1.1783961057662964,
+      "learning_rate": 2.385992285823596e-05,
+      "loss": 0.304,
+      "step": 250
+    },
+    {
+      "epoch": 1.0282828282828282,
+      "grad_norm": 1.255103349685669,
+      "learning_rate": 2.358845205133152e-05,
+      "loss": 0.261,
+      "step": 255
+    },
+    {
+      "epoch": 1.0484848484848486,
+      "grad_norm": 1.2307968139648438,
+      "learning_rate": 2.3312729889815915e-05,
+      "loss": 0.2305,
+      "step": 260
+    },
+    {
+      "epoch": 1.0686868686868687,
+      "grad_norm": 1.1091794967651367,
+      "learning_rate": 2.3032892858433515e-05,
+      "loss": 0.2787,
+      "step": 265
+    },
+    {
+      "epoch": 1.0888888888888888,
+      "grad_norm": 1.3384209871292114,
+      "learning_rate": 2.2749079478823224e-05,
+      "loss": 0.2514,
+      "step": 270
+    },
+    {
+      "epoch": 1.1090909090909091,
+      "grad_norm": 1.2037593126296997,
+      "learning_rate": 2.24614302409491e-05,
+      "loss": 0.2549,
+      "step": 275
+    },
+    {
+      "epoch": 1.1292929292929292,
+      "grad_norm": 1.3832809925079346,
+      "learning_rate": 2.217008753355671e-05,
+      "loss": 0.2399,
+      "step": 280
+    },
+    {
+      "epoch": 1.1494949494949496,
+      "grad_norm": 1.1958180665969849,
+      "learning_rate": 2.1875195573689476e-05,
+      "loss": 0.2385,
+      "step": 285
+    },
+    {
+      "epoch": 1.1696969696969697,
+      "grad_norm": 1.184637188911438,
+      "learning_rate": 2.157690033530009e-05,
+      "loss": 0.2446,
+      "step": 290
+    },
+    {
+      "epoch": 1.1898989898989898,
+      "grad_norm": 1.2521557807922363,
+      "learning_rate": 2.1275349476992168e-05,
+      "loss": 0.2004,
+      "step": 295
+    },
+    {
+      "epoch": 1.2101010101010101,
+      "grad_norm": 1.42829167842865,
+      "learning_rate": 2.0970692268928065e-05,
+      "loss": 0.2025,
+      "step": 300
+    },
+    {
+      "epoch": 1.2303030303030302,
+      "grad_norm": 1.340794563293457,
+      "learning_rate": 2.066307951893887e-05,
+      "loss": 0.2255,
+      "step": 305
+    },
+    {
+      "epoch": 1.2505050505050506,
+      "grad_norm": 1.0471365451812744,
+      "learning_rate": 2.0352663497873313e-05,
+      "loss": 0.215,
+      "step": 310
+    },
+    {
+      "epoch": 1.2707070707070707,
+      "grad_norm": 1.2765172719955444,
+      "learning_rate": 2.0039597864222423e-05,
+      "loss": 0.2072,
+      "step": 315
+    },
+    {
+      "epoch": 1.290909090909091,
+      "grad_norm": 1.1466255187988281,
+      "learning_rate": 1.9724037588057334e-05,
+      "loss": 0.2114,
+      "step": 320
+    },
+    {
+      "epoch": 1.3111111111111111,
+      "grad_norm": 1.097985863685608,
+      "learning_rate": 1.9406138874317777e-05,
+      "loss": 0.1759,
+      "step": 325
+    },
+    {
+      "epoch": 1.3313131313131312,
+      "grad_norm": 1.1392395496368408,
+      "learning_rate": 1.9086059085489393e-05,
+      "loss": 0.189,
+      "step": 330
+    },
+    {
+      "epoch": 1.3515151515151516,
+      "grad_norm": 0.9970096945762634,
+      "learning_rate": 1.876395666370798e-05,
+      "loss": 0.1959,
+      "step": 335
+    },
+    {
+      "epoch": 1.3717171717171717,
+      "grad_norm": 1.0721200704574585,
+      "learning_rate": 1.8439991052329318e-05,
+      "loss": 0.2098,
+      "step": 340
+    },
+    {
+      "epoch": 1.391919191919192,
+      "grad_norm": 1.1882052421569824,
+      "learning_rate": 1.8114322617003417e-05,
+      "loss": 0.1765,
+      "step": 345
+    },
+    {
+      "epoch": 1.412121212121212,
+      "grad_norm": 1.3733917474746704,
+      "learning_rate": 1.7787112566292195e-05,
+      "loss": 0.1929,
+      "step": 350
+    },
+    {
+      "epoch": 1.4323232323232324,
+      "grad_norm": 1.1271265745162964,
+      "learning_rate": 1.745852287186989e-05,
+      "loss": 0.1708,
+      "step": 355
+    },
+    {
+      "epoch": 1.4525252525252526,
+      "grad_norm": 1.110434889793396,
+      "learning_rate": 1.712871618834577e-05,
+      "loss": 0.1591,
+      "step": 360
+    },
+    {
+      "epoch": 1.4727272727272727,
+      "grad_norm": 1.1028157472610474,
+      "learning_rate": 1.6797855772748735e-05,
+      "loss": 0.1627,
+      "step": 365
+    },
+    {
+      "epoch": 1.492929292929293,
+      "grad_norm": 1.0171635150909424,
+      "learning_rate": 1.646610540371375e-05,
+      "loss": 0.164,
+      "step": 370
+    },
+    {
+      "epoch": 1.513131313131313,
+      "grad_norm": 1.2900023460388184,
+      "learning_rate": 1.613362930041009e-05,
+      "loss": 0.1612,
+      "step": 375
+    },
+    {
+      "epoch": 1.5333333333333332,
+      "grad_norm": 1.0427008867263794,
+      "learning_rate": 1.580059204125148e-05,
+      "loss": 0.1387,
+      "step": 380
+    },
+    {
+      "epoch": 1.5535353535353535,
+      "grad_norm": 1.256961703300476,
+      "learning_rate": 1.5467158482428396e-05,
+      "loss": 0.1465,
+      "step": 385
+    },
+    {
+      "epoch": 1.5737373737373739,
+      "grad_norm": 1.0609946250915527,
+      "learning_rate": 1.5133493676302943e-05,
+      "loss": 0.1651,
+      "step": 390
+    },
+    {
+      "epoch": 1.593939393939394,
+      "grad_norm": 1.1708347797393799,
+      "learning_rate": 1.4799762789706533e-05,
+      "loss": 0.138,
+      "step": 395
+    },
+    {
+      "epoch": 1.614141414141414,
+      "grad_norm": 0.9955141544342041,
+      "learning_rate": 1.4466131022180969e-05,
+      "loss": 0.1629,
+      "step": 400
+    },
+    {
+      "epoch": 1.6343434343434344,
+      "grad_norm": 0.9531594514846802,
+      "learning_rate": 1.4132763524203282e-05,
+      "loss": 0.126,
+      "step": 405
+    },
+    {
+      "epoch": 1.6545454545454545,
+      "grad_norm": 1.0088530778884888,
+      "learning_rate": 1.3799825315434915e-05,
+      "loss": 0.1202,
+      "step": 410
+    },
+    {
+      "epoch": 1.6747474747474747,
+      "grad_norm": 1.1042792797088623,
+      "learning_rate": 1.3467481203035596e-05,
+      "loss": 0.1345,
+      "step": 415
+    },
+    {
+      "epoch": 1.694949494949495,
+      "grad_norm": 1.0768200159072876,
+      "learning_rate": 1.3135895700082447e-05,
+      "loss": 0.1462,
+      "step": 420
+    },
+    {
+      "epoch": 1.7151515151515153,
+      "grad_norm": 1.1252286434173584,
+      "learning_rate": 1.280523294413467e-05,
+      "loss": 0.1347,
+      "step": 425
+    },
+    {
+      "epoch": 1.7353535353535352,
+      "grad_norm": 0.9585935473442078,
+      "learning_rate": 1.2475656615984087e-05,
+      "loss": 0.1198,
+      "step": 430
+    },
+    {
+      "epoch": 1.7555555555555555,
+      "grad_norm": 0.9740949869155884,
+      "learning_rate": 1.214732985863178e-05,
+      "loss": 0.1094,
+      "step": 435
+    },
+    {
+      "epoch": 1.7757575757575759,
+      "grad_norm": 1.2437893152236938,
+      "learning_rate": 1.1820415196531006e-05,
+      "loss": 0.1174,
+      "step": 440
+    },
+    {
+      "epoch": 1.795959595959596,
+      "grad_norm": 1.0259506702423096,
+      "learning_rate": 1.1495074455136192e-05,
+      "loss": 0.116,
+      "step": 445
+    },
+    {
+      "epoch": 1.816161616161616,
+      "grad_norm": 1.1177220344543457,
+      "learning_rate": 1.1171468680798063e-05,
+      "loss": 0.1163,
+      "step": 450
+    },
+    {
+      "epoch": 1.8363636363636364,
+      "grad_norm": 1.0267847776412964,
+      "learning_rate": 1.0849758061044325e-05,
+      "loss": 0.1265,
+      "step": 455
+    },
+    {
+      "epoch": 1.8565656565656565,
+      "grad_norm": 0.7422696948051453,
+      "learning_rate": 1.0530101845285524e-05,
+      "loss": 0.1048,
+      "step": 460
+    },
+    {
+      "epoch": 1.8767676767676766,
+      "grad_norm": 1.0799849033355713,
+      "learning_rate": 1.02126582659853e-05,
+      "loss": 0.1065,
+      "step": 465
+    },
+    {
+      "epoch": 1.896969696969697,
+      "grad_norm": 0.9203501343727112,
+      "learning_rate": 9.897584460333978e-06,
+      "loss": 0.1245,
+      "step": 470
+    },
+    {
+      "epoch": 1.9171717171717173,
+      "grad_norm": 1.1958577632904053,
+      "learning_rate": 9.585036392464382e-06,
+      "loss": 0.0985,
+      "step": 475
+    },
+    {
+      "epoch": 1.9373737373737374,
+      "grad_norm": 1.1301112174987793,
+      "learning_rate": 9.275168776248316e-06,
+      "loss": 0.112,
+      "step": 480
+    },
+    {
+      "epoch": 1.9575757575757575,
+      "grad_norm": 0.9478124976158142,
+      "learning_rate": 8.968134998711877e-06,
+      "loss": 0.1085,
+      "step": 485
+    },
+    {
+      "epoch": 1.9777777777777779,
+      "grad_norm": 1.0381243228912354,
+      "learning_rate": 8.664087044107664e-06,
+      "loss": 0.095,
+      "step": 490
+    },
+    {
+      "epoch": 1.997979797979798,
+      "grad_norm": 0.9924798607826233,
+      "learning_rate": 8.363175418681276e-06,
+      "loss": 0.1077,
+      "step": 495
+    },
+    {
+      "epoch": 2.0161616161616163,
+      "grad_norm": 1.0334126949310303,
+      "learning_rate": 8.065549076169506e-06,
+      "loss": 0.0744,
+      "step": 500
+    },
+    {
+      "epoch": 2.036363636363636,
+      "grad_norm": 0.8067806959152222,
+      "learning_rate": 7.77135534406701e-06,
+      "loss": 0.0736,
+      "step": 505
+    },
+    {
+      "epoch": 2.0565656565656565,
+      "grad_norm": 0.7703298926353455,
+      "learning_rate": 7.48073985069796e-06,
+      "loss": 0.0701,
+      "step": 510
+    },
+    {
+      "epoch": 2.076767676767677,
+      "grad_norm": 0.6393172144889832,
+      "learning_rate": 7.193846453128853e-06,
+      "loss": 0.0908,
+      "step": 515
+    },
+    {
+      "epoch": 2.096969696969697,
+      "grad_norm": 0.7883558869361877,
+      "learning_rate": 6.910817165958064e-06,
+      "loss": 0.0709,
+      "step": 520
+    },
+    {
+      "epoch": 2.117171717171717,
+      "grad_norm": 0.8169750571250916,
+      "learning_rate": 6.631792091017424e-06,
+      "loss": 0.0743,
+      "step": 525
+    },
+    {
+      "epoch": 2.1373737373737374,
+      "grad_norm": 0.6672986149787903,
+      "learning_rate": 6.356909348020708e-06,
+      "loss": 0.0745,
+      "step": 530
+    },
+    {
+      "epoch": 2.1575757575757577,
+      "grad_norm": 0.7301447987556458,
+      "learning_rate": 6.086305006193188e-06,
+      "loss": 0.0713,
+      "step": 535
+    },
+    {
+      "epoch": 2.1777777777777776,
+      "grad_norm": 0.8023566007614136,
+      "learning_rate": 5.820113016916266e-06,
+      "loss": 0.0701,
+      "step": 540
+    },
+    {
+      "epoch": 2.197979797979798,
+      "grad_norm": 0.6537421345710754,
+      "learning_rate": 5.55846514742045e-06,
+      "loss": 0.0644,
+      "step": 545
+    },
+    {
+      "epoch": 2.2181818181818183,
+      "grad_norm": 0.8120123147964478,
+      "learning_rate": 5.301490915559458e-06,
+      "loss": 0.0731,
+      "step": 550
+    },
+    {
+      "epoch": 2.2383838383838386,
+      "grad_norm": 0.7237561941146851,
+      "learning_rate": 5.049317525697824e-06,
+      "loss": 0.0597,
+      "step": 555
+    },
+    {
+      "epoch": 2.2585858585858585,
+      "grad_norm": 0.5892989635467529,
+      "learning_rate": 4.8020698057437066e-06,
+      "loss": 0.0671,
+      "step": 560
+    },
+    {
+      "epoch": 2.278787878787879,
+      "grad_norm": 0.7169414162635803,
+      "learning_rate": 4.559870145358008e-06,
+      "loss": 0.0702,
+      "step": 565
+    },
+    {
+      "epoch": 2.298989898989899,
+      "grad_norm": 0.659731924533844,
+      "learning_rate": 4.32283843537052e-06,
+      "loss": 0.0677,
+      "step": 570
+    },
+    {
+      "epoch": 2.319191919191919,
+      "grad_norm": 1.0039764642715454,
+      "learning_rate": 4.091092008432955e-06,
+      "loss": 0.0652,
+      "step": 575
+    },
+    {
+      "epoch": 2.3393939393939394,
+      "grad_norm": 0.7391262054443359,
+      "learning_rate": 3.864745580938301e-06,
+      "loss": 0.0628,
+      "step": 580
+    },
+    {
+      "epoch": 2.3595959595959597,
+      "grad_norm": 0.742656946182251,
+      "learning_rate": 3.643911196235302e-06,
+      "loss": 0.0722,
+      "step": 585
+    },
+    {
+      "epoch": 2.3797979797979796,
+      "grad_norm": 0.555101215839386,
+      "learning_rate": 3.4286981691660307e-06,
+      "loss": 0.0619,
+      "step": 590
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 0.5695873498916626,
+      "learning_rate": 3.2192130319541656e-06,
+      "loss": 0.0673,
+      "step": 595
+    },
+    {
+      "epoch": 2.4202020202020202,
+      "grad_norm": 0.6123603582382202,
+      "learning_rate": 3.0155594814706715e-06,
+      "loss": 0.0713,
+      "step": 600
+    },
+    {
+      "epoch": 2.4404040404040406,
+      "grad_norm": 0.6142969131469727,
+      "learning_rate": 2.8178383279029563e-06,
+      "loss": 0.0587,
+      "step": 605
+    },
+    {
+      "epoch": 2.4606060606060605,
+      "grad_norm": 0.5539048910140991,
+      "learning_rate": 2.6261474448530358e-06,
+      "loss": 0.0568,
+      "step": 610
+    },
+    {
+      "epoch": 2.480808080808081,
+      "grad_norm": 0.6815639138221741,
+      "learning_rate": 2.440581720889263e-06,
+      "loss": 0.0697,
+      "step": 615
+    },
+    {
+      "epoch": 2.501010101010101,
+      "grad_norm": 0.6519346833229065,
+      "learning_rate": 2.261233012575727e-06,
+      "loss": 0.0547,
+      "step": 620
+    },
+    {
+      "epoch": 2.5212121212121215,
+      "grad_norm": 0.6399024724960327,
+      "learning_rate": 2.0881900990024843e-06,
+      "loss": 0.058,
+      "step": 625
+    },
+    {
+      "epoch": 2.5414141414141413,
+      "grad_norm": 0.6960001587867737,
+      "learning_rate": 1.9215386378391763e-06,
+      "loss": 0.0658,
+      "step": 630
+    },
+    {
+      "epoch": 2.5616161616161617,
+      "grad_norm": 0.6924479603767395,
+      "learning_rate": 1.761361122933774e-06,
+      "loss": 0.067,
+      "step": 635
+    },
+    {
+      "epoch": 2.581818181818182,
+      "grad_norm": 0.6626688241958618,
+      "learning_rate": 1.6077368434774598e-06,
+      "loss": 0.0579,
+      "step": 640
+    },
+    {
+      "epoch": 2.602020202020202,
+      "grad_norm": 0.5447858572006226,
+      "learning_rate": 1.4607418447557973e-06,
+      "loss": 0.0616,
+      "step": 645
+    },
+    {
+      "epoch": 2.6222222222222222,
+      "grad_norm": 0.6896591186523438,
+      "learning_rate": 1.3204488905057161e-06,
+      "loss": 0.0525,
+      "step": 650
+    },
+    {
+      "epoch": 2.6424242424242426,
+      "grad_norm": 0.5828932523727417,
+      "learning_rate": 1.1869274268968355e-06,
+      "loss": 0.0584,
+      "step": 655
+    },
+    {
+      "epoch": 2.6626262626262625,
+      "grad_norm": 0.6555042862892151,
+      "learning_rate": 1.0602435481550532e-06,
+      "loss": 0.0587,
+      "step": 660
+    },
+    {
+      "epoch": 2.682828282828283,
+      "grad_norm": 0.6443304419517517,
+      "learning_rate": 9.404599638453598e-07,
+      "loss": 0.0609,
+      "step": 665
+    },
+    {
+      "epoch": 2.703030303030303,
+      "grad_norm": 0.5971073508262634,
+      "learning_rate": 8.276359678300754e-07,
+      "loss": 0.0532,
+      "step": 670
+    },
+    {
+      "epoch": 2.723232323232323,
+      "grad_norm": 0.6533417701721191,
+      "learning_rate": 7.218274089179094e-07,
+      "loss": 0.0568,
+      "step": 675
+    },
+    {
+      "epoch": 2.7434343434343433,
+      "grad_norm": 0.683943510055542,
+      "learning_rate": 6.230866632183346e-07,
+      "loss": 0.0618,
+      "step": 680
+    },
+    {
+      "epoch": 2.7636363636363637,
+      "grad_norm": 0.5820612907409668,
+      "learning_rate": 5.314626082149765e-07,
+      "loss": 0.0576,
+      "step": 685
+    },
+    {
+      "epoch": 2.783838383838384,
+      "grad_norm": 0.6201310753822327,
+      "learning_rate": 4.470005985708669e-07,
+      "loss": 0.0579,
+      "step": 690
+    },
+    {
+      "epoch": 2.804040404040404,
+      "grad_norm": 0.4896237850189209,
+      "learning_rate": 3.697424436775032e-07,
+      "loss": 0.0577,
+      "step": 695
+    },
+    {
+      "epoch": 2.824242424242424,
+      "grad_norm": 0.5821230411529541,
+      "learning_rate": 2.9972638695885455e-07,
+      "loss": 0.0629,
+      "step": 700
+    },
+    {
+      "epoch": 2.8444444444444446,
+      "grad_norm": 0.5690262913703918,
+      "learning_rate": 2.3698708694055883e-07,
+      "loss": 0.0579,
+      "step": 705
+    },
+    {
+      "epoch": 2.864646464646465,
+      "grad_norm": 0.4587281346321106,
+      "learning_rate": 1.8155560009365746e-07,
+      "loss": 0.0535,
+      "step": 710
+    },
+    {
+      "epoch": 2.8848484848484848,
+      "grad_norm": 0.6618624329566956,
+      "learning_rate": 1.334593654613897e-07,
+      "loss": 0.0607,
+      "step": 715
+    },
+    {
+      "epoch": 2.905050505050505,
+      "grad_norm": 0.5596798062324524,
+      "learning_rate": 9.272219107664403e-08,
+      "loss": 0.0628,
+      "step": 720
+    },
+    {
+      "epoch": 2.9252525252525254,
+      "grad_norm": 0.5585241913795471,
+      "learning_rate": 5.936424217678038e-08,
+      "loss": 0.0583,
+      "step": 725
+    },
+    {
+      "epoch": 2.9454545454545453,
+      "grad_norm": 0.5195132493972778,
+      "learning_rate": 3.340203122167152e-08,
+      "loss": 0.0557,
+      "step": 730
+    },
+    {
+      "epoch": 2.9656565656565657,
+      "grad_norm": 1.0016818046569824,
+      "learning_rate": 1.4848409719901423e-08,
+      "loss": 0.0563,
+      "step": 735
+    },
+    {
+      "epoch": 2.985858585858586,
+      "grad_norm": 0.5090447068214417,
+      "learning_rate": 3.7125618671624004e-09,
+      "loss": 0.0537,
+      "step": 740
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 744,
+  "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": 9.365904541561651e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

12_128_e3_3e-5/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47d84cdbf40290f7c77f0b0aebb9d75fc91fd51a54cf2730abfa000a9b128b26
+size 8209

12_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)