Upload folder using huggingface_hub

109_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": [
+    "gate_proj",
+    "v_proj",
+    "down_proj",
+    "up_proj",
+    "o_proj",
+    "q_proj",
+    "k_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

109_128_e3_3e-5/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17ce29bf373cb03f47458159376038818b10bc6bfaf72a4479b5e9471bd3d32c
+size 791751704

109_128_e3_3e-5/latest

@@ -0,0 +1 @@
+global_step741

109_128_e3_3e-5/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48eb68ffc94ddc9d1c75817870699005d6b06f3254134545c42134de629d9983
+size 16389

109_128_e3_3e-5/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0151e70f64327e17a2efe828fe197276ba0b78b48626301b6f05cba57bd10943
+size 16389

109_128_e3_3e-5/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fb00cf053a423bfd7b0cd3811204b016a55186bc97612d3a676d815ecb9d446
+size 16389

109_128_e3_3e-5/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d6f2f5b8e6e2b44b3509820cbcaed46f5c263faa24b1de5978df18f321702378
+size 16389

109_128_e3_3e-5/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d6013fe063eb64748768aae287cb14145954f32c40a0fdf7fe363ccb0ad4eb1e
+size 16389

109_128_e3_3e-5/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:769cb4eeefb4cd0338e30508cfb39edb8638aba46b4f318e1d063e4350e07118
+size 16389

109_128_e3_3e-5/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac29277bc7d12b8e645a7c8dc3aa31ed0939578af59d52825c43774d08308c83
+size 16389

109_128_e3_3e-5/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc301586615b3c3a5286b191205b5e34ae6429ebbc162a966f19d6363af214fd
+size 16389

109_128_e3_3e-5/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:674c5775be9c37705ceaa3fdcf4515cea3fcd318a423bfc63281956e52c27ab3
+size 1401

109_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
+  }
+}

109_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
+}

109_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": 741,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.020242914979757085,
+      "grad_norm": 1.004509449005127,
+      "learning_rate": 3.157894736842105e-06,
+      "loss": 1.4884,
+      "step": 5
+    },
+    {
+      "epoch": 0.04048582995951417,
+      "grad_norm": 0.68101567029953,
+      "learning_rate": 7.105263157894737e-06,
+      "loss": 1.4283,
+      "step": 10
+    },
+    {
+      "epoch": 0.06072874493927125,
+      "grad_norm": 0.6341578960418701,
+      "learning_rate": 1.1052631578947368e-05,
+      "loss": 1.3918,
+      "step": 15
+    },
+    {
+      "epoch": 0.08097165991902834,
+      "grad_norm": 0.5399417877197266,
+      "learning_rate": 1.5e-05,
+      "loss": 1.3547,
+      "step": 20
+    },
+    {
+      "epoch": 0.10121457489878542,
+      "grad_norm": 0.5037768483161926,
+      "learning_rate": 1.894736842105263e-05,
+      "loss": 1.3281,
+      "step": 25
+    },
+    {
+      "epoch": 0.1214574898785425,
+      "grad_norm": 0.5652232766151428,
+      "learning_rate": 2.2894736842105263e-05,
+      "loss": 1.2656,
+      "step": 30
+    },
+    {
+      "epoch": 0.1417004048582996,
+      "grad_norm": 0.5854511260986328,
+      "learning_rate": 2.6842105263157896e-05,
+      "loss": 1.3529,
+      "step": 35
+    },
+    {
+      "epoch": 0.16194331983805668,
+      "grad_norm": 0.5356202125549316,
+      "learning_rate": 2.999985022144514e-05,
+      "loss": 1.2484,
+      "step": 40
+    },
+    {
+      "epoch": 0.18218623481781376,
+      "grad_norm": 0.6571934819221497,
+      "learning_rate": 2.9994608286088872e-05,
+      "loss": 1.224,
+      "step": 45
+    },
+    {
+      "epoch": 0.20242914979757085,
+      "grad_norm": 0.7864879965782166,
+      "learning_rate": 2.9981880413869707e-05,
+      "loss": 1.1521,
+      "step": 50
+    },
+    {
+      "epoch": 0.22267206477732793,
+      "grad_norm": 0.6652578711509705,
+      "learning_rate": 2.9961672959074872e-05,
+      "loss": 1.2023,
+      "step": 55
+    },
+    {
+      "epoch": 0.242914979757085,
+      "grad_norm": 0.7305110692977905,
+      "learning_rate": 2.993399601011266e-05,
+      "loss": 1.1317,
+      "step": 60
+    },
+    {
+      "epoch": 0.2631578947368421,
+      "grad_norm": 0.9022772908210754,
+      "learning_rate": 2.9898863384475917e-05,
+      "loss": 1.0915,
+      "step": 65
+    },
+    {
+      "epoch": 0.2834008097165992,
+      "grad_norm": 0.7142044901847839,
+      "learning_rate": 2.985629262184371e-05,
+      "loss": 1.0583,
+      "step": 70
+    },
+    {
+      "epoch": 0.30364372469635625,
+      "grad_norm": 0.7476593852043152,
+      "learning_rate": 2.9806304975324853e-05,
+      "loss": 1.0348,
+      "step": 75
+    },
+    {
+      "epoch": 0.32388663967611336,
+      "grad_norm": 0.8483200073242188,
+      "learning_rate": 2.97489254008474e-05,
+      "loss": 1.0121,
+      "step": 80
+    },
+    {
+      "epoch": 0.3441295546558704,
+      "grad_norm": 0.8936290740966797,
+      "learning_rate": 2.9684182544699645e-05,
+      "loss": 0.9573,
+      "step": 85
+    },
+    {
+      "epoch": 0.3643724696356275,
+      "grad_norm": 0.921067476272583,
+      "learning_rate": 2.9612108729228678e-05,
+      "loss": 0.9168,
+      "step": 90
+    },
+    {
+      "epoch": 0.38461538461538464,
+      "grad_norm": 0.984555184841156,
+      "learning_rate": 2.9532739936703703e-05,
+      "loss": 0.9142,
+      "step": 95
+    },
+    {
+      "epoch": 0.4048582995951417,
+      "grad_norm": 1.0167465209960938,
+      "learning_rate": 2.9446115791352207e-05,
+      "loss": 0.8921,
+      "step": 100
+    },
+    {
+      "epoch": 0.4251012145748988,
+      "grad_norm": 0.9743611812591553,
+      "learning_rate": 2.935227953957787e-05,
+      "loss": 0.8729,
+      "step": 105
+    },
+    {
+      "epoch": 0.44534412955465585,
+      "grad_norm": 0.9121851325035095,
+      "learning_rate": 2.9251278028370137e-05,
+      "loss": 0.8774,
+      "step": 110
+    },
+    {
+      "epoch": 0.46558704453441296,
+      "grad_norm": 0.9452496767044067,
+      "learning_rate": 2.9143161681916264e-05,
+      "loss": 0.8319,
+      "step": 115
+    },
+    {
+      "epoch": 0.48582995951417,
+      "grad_norm": 1.0705139636993408,
+      "learning_rate": 2.9027984476427427e-05,
+      "loss": 0.7785,
+      "step": 120
+    },
+    {
+      "epoch": 0.5060728744939271,
+      "grad_norm": 1.1426186561584473,
+      "learning_rate": 2.8905803913191545e-05,
+      "loss": 0.7181,
+      "step": 125
+    },
+    {
+      "epoch": 0.5263157894736842,
+      "grad_norm": 1.7189124822616577,
+      "learning_rate": 2.877668098986624e-05,
+      "loss": 0.6951,
+      "step": 130
+    },
+    {
+      "epoch": 0.5465587044534413,
+      "grad_norm": 1.4203096628189087,
+      "learning_rate": 2.864068017002621e-05,
+      "loss": 0.7745,
+      "step": 135
+    },
+    {
+      "epoch": 0.5668016194331984,
+      "grad_norm": 1.099601149559021,
+      "learning_rate": 2.8497869350980376e-05,
+      "loss": 0.6912,
+      "step": 140
+    },
+    {
+      "epoch": 0.5870445344129555,
+      "grad_norm": 1.0917069911956787,
+      "learning_rate": 2.834831982987462e-05,
+      "loss": 0.6874,
+      "step": 145
+    },
+    {
+      "epoch": 0.6072874493927125,
+      "grad_norm": 1.0699445009231567,
+      "learning_rate": 2.8192106268097336e-05,
+      "loss": 0.6651,
+      "step": 150
+    },
+    {
+      "epoch": 0.6275303643724697,
+      "grad_norm": 1.039550542831421,
+      "learning_rate": 2.8029306654005288e-05,
+      "loss": 0.6437,
+      "step": 155
+    },
+    {
+      "epoch": 0.6477732793522267,
+      "grad_norm": 1.194151759147644,
+      "learning_rate": 2.7860002263988565e-05,
+      "loss": 0.5713,
+      "step": 160
+    },
+    {
+      "epoch": 0.6680161943319838,
+      "grad_norm": 1.2917087078094482,
+      "learning_rate": 2.768427762189402e-05,
+      "loss": 0.645,
+      "step": 165
+    },
+    {
+      "epoch": 0.6882591093117408,
+      "grad_norm": 1.1495329141616821,
+      "learning_rate": 2.75022204568274e-05,
+      "loss": 0.5835,
+      "step": 170
+    },
+    {
+      "epoch": 0.708502024291498,
+      "grad_norm": 1.6731566190719604,
+      "learning_rate": 2.7313921659355305e-05,
+      "loss": 0.5887,
+      "step": 175
+    },
+    {
+      "epoch": 0.728744939271255,
+      "grad_norm": 1.230734944343567,
+      "learning_rate": 2.7119475236128816e-05,
+      "loss": 0.5224,
+      "step": 180
+    },
+    {
+      "epoch": 0.7489878542510121,
+      "grad_norm": 1.2216949462890625,
+      "learning_rate": 2.691897826295143e-05,
+      "loss": 0.5156,
+      "step": 185
+    },
+    {
+      "epoch": 0.7692307692307693,
+      "grad_norm": 1.2396891117095947,
+      "learning_rate": 2.6712530836314763e-05,
+      "loss": 0.5417,
+      "step": 190
+    },
+    {
+      "epoch": 0.7894736842105263,
+      "grad_norm": 1.3565627336502075,
+      "learning_rate": 2.650023602342614e-05,
+      "loss": 0.4906,
+      "step": 195
+    },
+    {
+      "epoch": 0.8097165991902834,
+      "grad_norm": 1.5496858358383179,
+      "learning_rate": 2.6282199810753192e-05,
+      "loss": 0.5261,
+      "step": 200
+    },
+    {
+      "epoch": 0.8299595141700404,
+      "grad_norm": 1.3620166778564453,
+      "learning_rate": 2.6058531051110933e-05,
+      "loss": 0.5092,
+      "step": 205
+    },
+    {
+      "epoch": 0.8502024291497976,
+      "grad_norm": 1.4385701417922974,
+      "learning_rate": 2.5829341409317866e-05,
+      "loss": 0.448,
+      "step": 210
+    },
+    {
+      "epoch": 0.8704453441295547,
+      "grad_norm": 1.3609861135482788,
+      "learning_rate": 2.559474530644825e-05,
+      "loss": 0.4659,
+      "step": 215
+    },
+    {
+      "epoch": 0.8906882591093117,
+      "grad_norm": 1.3215820789337158,
+      "learning_rate": 2.5354859862708276e-05,
+      "loss": 0.4564,
+      "step": 220
+    },
+    {
+      "epoch": 0.9109311740890689,
+      "grad_norm": 1.351797103881836,
+      "learning_rate": 2.5109804838964785e-05,
+      "loss": 0.4753,
+      "step": 225
+    },
+    {
+      "epoch": 0.9311740890688259,
+      "grad_norm": 1.336039423942566,
+      "learning_rate": 2.4859702576955595e-05,
+      "loss": 0.4325,
+      "step": 230
+    },
+    {
+      "epoch": 0.951417004048583,
+      "grad_norm": 1.535072922706604,
+      "learning_rate": 2.4604677938211394e-05,
+      "loss": 0.4153,
+      "step": 235
+    },
+    {
+      "epoch": 0.97165991902834,
+      "grad_norm": 1.3180216550827026,
+      "learning_rate": 2.4344858241719625e-05,
+      "loss": 0.4161,
+      "step": 240
+    },
+    {
+      "epoch": 0.9919028340080972,
+      "grad_norm": 1.5556411743164062,
+      "learning_rate": 2.4080373200361527e-05,
+      "loss": 0.3806,
+      "step": 245
+    },
+    {
+      "epoch": 1.0121457489878543,
+      "grad_norm": 1.2471015453338623,
+      "learning_rate": 2.3811354856153996e-05,
+      "loss": 0.3743,
+      "step": 250
+    },
+    {
+      "epoch": 1.0323886639676114,
+      "grad_norm": 1.3908895254135132,
+      "learning_rate": 2.3537937514328704e-05,
+      "loss": 0.3279,
+      "step": 255
+    },
+    {
+      "epoch": 1.0526315789473684,
+      "grad_norm": 1.2971608638763428,
+      "learning_rate": 2.326025767628129e-05,
+      "loss": 0.2888,
+      "step": 260
+    },
+    {
+      "epoch": 1.0728744939271255,
+      "grad_norm": 1.7477117776870728,
+      "learning_rate": 2.297845397142415e-05,
+      "loss": 0.3192,
+      "step": 265
+    },
+    {
+      "epoch": 1.0931174089068827,
+      "grad_norm": 1.464966058731079,
+      "learning_rate": 2.2692667087976803e-05,
+      "loss": 0.3323,
+      "step": 270
+    },
+    {
+      "epoch": 1.1133603238866396,
+      "grad_norm": 1.5718823671340942,
+      "learning_rate": 2.2403039702728466e-05,
+      "loss": 0.277,
+      "step": 275
+    },
+    {
+      "epoch": 1.1336032388663968,
+      "grad_norm": 1.3580231666564941,
+      "learning_rate": 2.2109716409807804e-05,
+      "loss": 0.3076,
+      "step": 280
+    },
+    {
+      "epoch": 1.1538461538461537,
+      "grad_norm": 1.425230622291565,
+      "learning_rate": 2.1812843648495465e-05,
+      "loss": 0.2777,
+      "step": 285
+    },
+    {
+      "epoch": 1.174089068825911,
+      "grad_norm": 1.2472330331802368,
+      "learning_rate": 2.1512569630115507e-05,
+      "loss": 0.2552,
+      "step": 290
+    },
+    {
+      "epoch": 1.194331983805668,
+      "grad_norm": 1.3370387554168701,
+      "learning_rate": 2.1209044264042016e-05,
+      "loss": 0.3156,
+      "step": 295
+    },
+    {
+      "epoch": 1.214574898785425,
+      "grad_norm": 1.490308165550232,
+      "learning_rate": 2.0902419082858142e-05,
+      "loss": 0.2697,
+      "step": 300
+    },
+    {
+      "epoch": 1.2348178137651822,
+      "grad_norm": 1.4427396059036255,
+      "learning_rate": 2.059284716670463e-05,
+      "loss": 0.2564,
+      "step": 305
+    },
+    {
+      "epoch": 1.2550607287449393,
+      "grad_norm": 1.8000982999801636,
+      "learning_rate": 2.0280483066855823e-05,
+      "loss": 0.286,
+      "step": 310
+    },
+    {
+      "epoch": 1.2753036437246963,
+      "grad_norm": 1.2955548763275146,
+      "learning_rate": 1.996548272856117e-05,
+      "loss": 0.2619,
+      "step": 315
+    },
+    {
+      "epoch": 1.2955465587044535,
+      "grad_norm": 1.445777177810669,
+      "learning_rate": 1.9648003413190858e-05,
+      "loss": 0.2333,
+      "step": 320
+    },
+    {
+      "epoch": 1.3157894736842106,
+      "grad_norm": 1.342530369758606,
+      "learning_rate": 1.9328203619724297e-05,
+      "loss": 0.2201,
+      "step": 325
+    },
+    {
+      "epoch": 1.3360323886639676,
+      "grad_norm": 1.491340160369873,
+      "learning_rate": 1.9006243005620825e-05,
+      "loss": 0.274,
+      "step": 330
+    },
+    {
+      "epoch": 1.3562753036437247,
+      "grad_norm": 1.793227195739746,
+      "learning_rate": 1.8682282307111988e-05,
+      "loss": 0.2243,
+      "step": 335
+    },
+    {
+      "epoch": 1.376518218623482,
+      "grad_norm": 1.4045292139053345,
+      "learning_rate": 1.8356483258955302e-05,
+      "loss": 0.246,
+      "step": 340
+    },
+    {
+      "epoch": 1.3967611336032388,
+      "grad_norm": 1.431766152381897,
+      "learning_rate": 1.802900851368946e-05,
+      "loss": 0.2547,
+      "step": 345
+    },
+    {
+      "epoch": 1.417004048582996,
+      "grad_norm": 1.2871906757354736,
+      "learning_rate": 1.770002156043138e-05,
+      "loss": 0.2294,
+      "step": 350
+    },
+    {
+      "epoch": 1.4372469635627532,
+      "grad_norm": 1.2669167518615723,
+      "learning_rate": 1.73696866432556e-05,
+      "loss": 0.2269,
+      "step": 355
+    },
+    {
+      "epoch": 1.45748987854251,
+      "grad_norm": 2.1537318229675293,
+      "learning_rate": 1.70381686791967e-05,
+      "loss": 0.2417,
+      "step": 360
+    },
+    {
+      "epoch": 1.4777327935222673,
+      "grad_norm": 1.364776849746704,
+      "learning_rate": 1.670563317591589e-05,
+      "loss": 0.2076,
+      "step": 365
+    },
+    {
+      "epoch": 1.4979757085020242,
+      "grad_norm": 1.3802754878997803,
+      "learning_rate": 1.6372246149072572e-05,
+      "loss": 0.2125,
+      "step": 370
+    },
+    {
+      "epoch": 1.5182186234817814,
+      "grad_norm": 1.3760582208633423,
+      "learning_rate": 1.6038174039442408e-05,
+      "loss": 0.2021,
+      "step": 375
+    },
+    {
+      "epoch": 1.5384615384615383,
+      "grad_norm": 1.3496967554092407,
+      "learning_rate": 1.570358362982309e-05,
+      "loss": 0.1991,
+      "step": 380
+    },
+    {
+      "epoch": 1.5587044534412957,
+      "grad_norm": 1.7255041599273682,
+      "learning_rate": 1.5368641961769404e-05,
+      "loss": 0.1954,
+      "step": 385
+    },
+    {
+      "epoch": 1.5789473684210527,
+      "grad_norm": 1.186657190322876,
+      "learning_rate": 1.5033516252199047e-05,
+      "loss": 0.1939,
+      "step": 390
+    },
+    {
+      "epoch": 1.5991902834008096,
+      "grad_norm": 1.1751497983932495,
+      "learning_rate": 1.4698373809910964e-05,
+      "loss": 0.1709,
+      "step": 395
+    },
+    {
+      "epoch": 1.6194331983805668,
+      "grad_norm": 1.4781550168991089,
+      "learning_rate": 1.4363381952057779e-05,
+      "loss": 0.1975,
+      "step": 400
+    },
+    {
+      "epoch": 1.639676113360324,
+      "grad_norm": 1.571811318397522,
+      "learning_rate": 1.402870792061404e-05,
+      "loss": 0.2045,
+      "step": 405
+    },
+    {
+      "epoch": 1.6599190283400809,
+      "grad_norm": 1.2236011028289795,
+      "learning_rate": 1.3694518798882041e-05,
+      "loss": 0.1837,
+      "step": 410
+    },
+    {
+      "epoch": 1.680161943319838,
+      "grad_norm": 1.1486406326293945,
+      "learning_rate": 1.3360981428076837e-05,
+      "loss": 0.1723,
+      "step": 415
+    },
+    {
+      "epoch": 1.7004048582995952,
+      "grad_norm": 1.3695476055145264,
+      "learning_rate": 1.3028262324032083e-05,
+      "loss": 0.1871,
+      "step": 420
+    },
+    {
+      "epoch": 1.7206477732793521,
+      "grad_norm": 1.2928073406219482,
+      "learning_rate": 1.269652759406839e-05,
+      "loss": 0.1657,
+      "step": 425
+    },
+    {
+      "epoch": 1.7408906882591093,
+      "grad_norm": 1.1803869009017944,
+      "learning_rate": 1.236594285406557e-05,
+      "loss": 0.1561,
+      "step": 430
+    },
+    {
+      "epoch": 1.7611336032388665,
+      "grad_norm": 1.1030513048171997,
+      "learning_rate": 1.2036673145780251e-05,
+      "loss": 0.1378,
+      "step": 435
+    },
+    {
+      "epoch": 1.7813765182186234,
+      "grad_norm": 1.1215111017227173,
+      "learning_rate": 1.1708882854450123e-05,
+      "loss": 0.1785,
+      "step": 440
+    },
+    {
+      "epoch": 1.8016194331983806,
+      "grad_norm": 1.2139201164245605,
+      "learning_rate": 1.1382735626725916e-05,
+      "loss": 0.164,
+      "step": 445
+    },
+    {
+      "epoch": 1.8218623481781377,
+      "grad_norm": 1.1551944017410278,
+      "learning_rate": 1.105839428897216e-05,
+      "loss": 0.1443,
+      "step": 450
+    },
+    {
+      "epoch": 1.8421052631578947,
+      "grad_norm": 1.381704330444336,
+      "learning_rate": 1.0736020765977434e-05,
+      "loss": 0.1502,
+      "step": 455
+    },
+    {
+      "epoch": 1.8623481781376519,
+      "grad_norm": 0.9443936944007874,
+      "learning_rate": 1.0415776000114716e-05,
+      "loss": 0.1311,
+      "step": 460
+    },
+    {
+      "epoch": 1.882591093117409,
+      "grad_norm": 1.4079439640045166,
+      "learning_rate": 1.0097819870992196e-05,
+      "loss": 0.1434,
+      "step": 465
+    },
+    {
+      "epoch": 1.902834008097166,
+      "grad_norm": 1.212890863418579,
+      "learning_rate": 9.782311115634706e-06,
+      "loss": 0.1489,
+      "step": 470
+    },
+    {
+      "epoch": 1.9230769230769231,
+      "grad_norm": 1.4142554998397827,
+      "learning_rate": 9.469407249235532e-06,
+      "loss": 0.13,
+      "step": 475
+    },
+    {
+      "epoch": 1.9433198380566803,
+      "grad_norm": 1.0147905349731445,
+      "learning_rate": 9.1592644865182e-06,
+      "loss": 0.1402,
+      "step": 480
+    },
+    {
+      "epoch": 1.9635627530364372,
+      "grad_norm": 1.3028090000152588,
+      "learning_rate": 8.852037663747595e-06,
+      "loss": 0.1248,
+      "step": 485
+    },
+    {
+      "epoch": 1.9838056680161942,
+      "grad_norm": 1.25118887424469,
+      "learning_rate": 8.54788016142915e-06,
+      "loss": 0.1133,
+      "step": 490
+    },
+    {
+      "epoch": 2.0040485829959516,
+      "grad_norm": 1.0490270853042603,
+      "learning_rate": 8.246943827734899e-06,
+      "loss": 0.1154,
+      "step": 495
+    },
+    {
+      "epoch": 2.0242914979757085,
+      "grad_norm": 0.9214974641799927,
+      "learning_rate": 7.94937890269453e-06,
+      "loss": 0.0768,
+      "step": 500
+    },
+    {
+      "epoch": 2.0445344129554655,
+      "grad_norm": 1.0407683849334717,
+      "learning_rate": 7.655333943189218e-06,
+      "loss": 0.1061,
+      "step": 505
+    },
+    {
+      "epoch": 2.064777327935223,
+      "grad_norm": 0.985413134098053,
+      "learning_rate": 7.364955748785817e-06,
+      "loss": 0.1019,
+      "step": 510
+    },
+    {
+      "epoch": 2.08502024291498,
+      "grad_norm": 1.0390602350234985,
+      "learning_rate": 7.078389288448384e-06,
+      "loss": 0.0915,
+      "step": 515
+    },
+    {
+      "epoch": 2.1052631578947367,
+      "grad_norm": 1.145819067955017,
+      "learning_rate": 6.795777628163599e-06,
+      "loss": 0.0979,
+      "step": 520
+    },
+    {
+      "epoch": 2.125506072874494,
+      "grad_norm": 1.1324427127838135,
+      "learning_rate": 6.517261859516188e-06,
+      "loss": 0.0882,
+      "step": 525
+    },
+    {
+      "epoch": 2.145748987854251,
+      "grad_norm": 1.0639681816101074,
+      "learning_rate": 6.242981029250131e-06,
+      "loss": 0.0928,
+      "step": 530
+    },
+    {
+      "epoch": 2.165991902834008,
+      "grad_norm": 1.015256404876709,
+      "learning_rate": 5.973072069850686e-06,
+      "loss": 0.0968,
+      "step": 535
+    },
+    {
+      "epoch": 2.1862348178137654,
+      "grad_norm": 1.073889136314392,
+      "learning_rate": 5.707669731181955e-06,
+      "loss": 0.0844,
+      "step": 540
+    },
+    {
+      "epoch": 2.2064777327935223,
+      "grad_norm": 1.00910222530365,
+      "learning_rate": 5.4469065132141025e-06,
+      "loss": 0.0837,
+      "step": 545
+    },
+    {
+      "epoch": 2.2267206477732793,
+      "grad_norm": 1.0402616262435913,
+      "learning_rate": 5.190912599873818e-06,
+      "loss": 0.0971,
+      "step": 550
+    },
+    {
+      "epoch": 2.246963562753036,
+      "grad_norm": 0.8678575158119202,
+      "learning_rate": 4.93981579405105e-06,
+      "loss": 0.0941,
+      "step": 555
+    },
+    {
+      "epoch": 2.2672064777327936,
+      "grad_norm": 1.2524758577346802,
+      "learning_rate": 4.693741453794433e-06,
+      "loss": 0.089,
+      "step": 560
+    },
+    {
+      "epoch": 2.2874493927125505,
+      "grad_norm": 0.8918817043304443,
+      "learning_rate": 4.452812429727313e-06,
+      "loss": 0.0924,
+      "step": 565
+    },
+    {
+      "epoch": 2.3076923076923075,
+      "grad_norm": 0.9283481240272522,
+      "learning_rate": 4.21714900371556e-06,
+      "loss": 0.0871,
+      "step": 570
+    },
+    {
+      "epoch": 2.327935222672065,
+      "grad_norm": 0.9137043952941895,
+      "learning_rate": 3.986868828817818e-06,
+      "loss": 0.0941,
+      "step": 575
+    },
+    {
+      "epoch": 2.348178137651822,
+      "grad_norm": 0.8159996867179871,
+      "learning_rate": 3.7620868705481586e-06,
+      "loss": 0.0798,
+      "step": 580
+    },
+    {
+      "epoch": 2.3684210526315788,
+      "grad_norm": 0.9702566862106323,
+      "learning_rate": 3.5429153494805087e-06,
+      "loss": 0.0869,
+      "step": 585
+    },
+    {
+      "epoch": 2.388663967611336,
+      "grad_norm": 1.0760486125946045,
+      "learning_rate": 3.329463685223411e-06,
+      "loss": 0.0806,
+      "step": 590
+    },
+    {
+      "epoch": 2.408906882591093,
+      "grad_norm": 0.9018325805664062,
+      "learning_rate": 3.12183844179316e-06,
+      "loss": 0.0767,
+      "step": 595
+    },
+    {
+      "epoch": 2.42914979757085,
+      "grad_norm": 0.8729963302612305,
+      "learning_rate": 2.9201432744126074e-06,
+      "loss": 0.0734,
+      "step": 600
+    },
+    {
+      "epoch": 2.4493927125506074,
+      "grad_norm": 0.8886759877204895,
+      "learning_rate": 2.7244788777621326e-06,
+      "loss": 0.0782,
+      "step": 605
+    },
+    {
+      "epoch": 2.4696356275303644,
+      "grad_norm": 0.8715010285377502,
+      "learning_rate": 2.5349429357085952e-06,
+      "loss": 0.0768,
+      "step": 610
+    },
+    {
+      "epoch": 2.4898785425101213,
+      "grad_norm": 1.174996256828308,
+      "learning_rate": 2.3516300725375016e-06,
+      "loss": 0.0737,
+      "step": 615
+    },
+    {
+      "epoch": 2.5101214574898787,
+      "grad_norm": 0.8888649344444275,
+      "learning_rate": 2.174631805712576e-06,
+      "loss": 0.0809,
+      "step": 620
+    },
+    {
+      "epoch": 2.5303643724696356,
+      "grad_norm": 0.8588908314704895,
+      "learning_rate": 2.0040365001864074e-06,
+      "loss": 0.0655,
+      "step": 625
+    },
+    {
+      "epoch": 2.5506072874493926,
+      "grad_norm": 0.8345527648925781,
+      "learning_rate": 1.8399293242849635e-06,
+      "loss": 0.0713,
+      "step": 630
+    },
+    {
+      "epoch": 2.57085020242915,
+      "grad_norm": 1.0172175168991089,
+      "learning_rate": 1.6823922071879932e-06,
+      "loss": 0.0869,
+      "step": 635
+    },
+    {
+      "epoch": 2.591093117408907,
+      "grad_norm": 0.869384765625,
+      "learning_rate": 1.5315037980265478e-06,
+      "loss": 0.0692,
+      "step": 640
+    },
+    {
+      "epoch": 2.611336032388664,
+      "grad_norm": 0.7183501720428467,
+      "learning_rate": 1.3873394266180377e-06,
+      "loss": 0.0702,
+      "step": 645
+    },
+    {
+      "epoch": 2.6315789473684212,
+      "grad_norm": 0.8675611019134521,
+      "learning_rate": 1.2499710658584296e-06,
+      "loss": 0.0753,
+      "step": 650
+    },
+    {
+      "epoch": 2.651821862348178,
+      "grad_norm": 0.8756380677223206,
+      "learning_rate": 1.119467295790363e-06,
+      "loss": 0.0658,
+      "step": 655
+    },
+    {
+      "epoch": 2.672064777327935,
+      "grad_norm": 1.0383371114730835,
+      "learning_rate": 9.958932693651168e-07,
+      "loss": 0.077,
+      "step": 660
+    },
+    {
+      "epoch": 2.6923076923076925,
+      "grad_norm": 0.843620777130127,
+      "learning_rate": 8.793106799155187e-07,
+      "loss": 0.0685,
+      "step": 665
+    },
+    {
+      "epoch": 2.7125506072874495,
+      "grad_norm": 0.6814716458320618,
+      "learning_rate": 7.697777303560699e-07,
+      "loss": 0.0595,
+      "step": 670
+    },
+    {
+      "epoch": 2.7327935222672064,
+      "grad_norm": 0.8059114813804626,
+      "learning_rate": 6.673491041255874e-07,
+      "loss": 0.0731,
+      "step": 675
+    },
+    {
+      "epoch": 2.753036437246964,
+      "grad_norm": 0.7944269776344299,
+      "learning_rate": 5.720759378869455e-07,
+      "loss": 0.0791,
+      "step": 680
+    },
+    {
+      "epoch": 2.7732793522267207,
+      "grad_norm": 0.6896971464157104,
+      "learning_rate": 4.840057959975169e-07,
+      "loss": 0.0673,
+      "step": 685
+    },
+    {
+      "epoch": 2.7935222672064777,
+      "grad_norm": 0.7642562985420227,
+      "learning_rate": 4.031826467630545e-07,
+      "loss": 0.0633,
+      "step": 690
+    },
+    {
+      "epoch": 2.813765182186235,
+      "grad_norm": 0.7817161083221436,
+      "learning_rate": 3.2964684048685523e-07,
+      "loss": 0.0715,
+      "step": 695
+    },
+    {
+      "epoch": 2.834008097165992,
+      "grad_norm": 0.8521256446838379,
+      "learning_rate": 2.6343508932521243e-07,
+      "loss": 0.0634,
+      "step": 700
+    },
+    {
+      "epoch": 2.854251012145749,
+      "grad_norm": 0.9094536304473877,
+      "learning_rate": 2.0458044895916516e-07,
+      "loss": 0.0735,
+      "step": 705
+    },
+    {
+      "epoch": 2.8744939271255063,
+      "grad_norm": 0.7276894450187683,
+      "learning_rate": 1.5311230209171078e-07,
+      "loss": 0.0722,
+      "step": 710
+    },
+    {
+      "epoch": 2.8947368421052633,
+      "grad_norm": 0.8718521595001221,
+      "learning_rate": 1.0905634377873563e-07,
+      "loss": 0.0747,
+      "step": 715
+    },
+    {
+      "epoch": 2.91497975708502,
+      "grad_norm": 0.780195415019989,
+      "learning_rate": 7.243456860096476e-08,
+      "loss": 0.0722,
+      "step": 720
+    },
+    {
+      "epoch": 2.9352226720647776,
+      "grad_norm": 0.8221401572227478,
+      "learning_rate": 4.326525968334216e-08,
+      "loss": 0.0699,
+      "step": 725
+    },
+    {
+      "epoch": 2.9554655870445345,
+      "grad_norm": 0.5896940231323242,
+      "learning_rate": 2.1562979567330554e-08,
+      "loss": 0.0741,
+      "step": 730
+    },
+    {
+      "epoch": 2.9757085020242915,
+      "grad_norm": 0.800220787525177,
+      "learning_rate": 7.338562940680249e-09,
+      "loss": 0.0669,
+      "step": 735
+    },
+    {
+      "epoch": 2.9959514170040484,
+      "grad_norm": 0.8188586235046387,
+      "learning_rate": 5.991112283026956e-10,
+      "loss": 0.0702,
+      "step": 740
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 741,
+  "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": 9.59451742586536e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

109_128_e3_3e-5/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c094d55b3e622cd94fec351550b2f61b81a6d7aa8b37cc46d16625c34d8abe02
+size 8209

109_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)