Upload folder using huggingface_hub

118_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": [
+    "k_proj",
+    "q_proj",
+    "up_proj",
+    "o_proj",
+    "down_proj",
+    "gate_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

118_128_e3_3e-5/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f0f66c178c3550d791fa1e3015d041294954f0bb966d32c02fbc383b83f6ca8
+size 791751704

118_128_e3_3e-5/latest

@@ -0,0 +1 @@
+global_step585

118_128_e3_3e-5/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d90303b62d7e9d64260331286bcdbea9c8cfcad841b340064d5cf7aae3c4cb75
+size 16389

118_128_e3_3e-5/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2be2c0b198a0753ac434d904fb15d2672bbbbdc6d6573f41b1fbb333e1d85fea
+size 16389

118_128_e3_3e-5/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f1365b08611d3a2fbec759cba47d12c5d82d37e83958ae8b720088c58bd5dc69
+size 16389

118_128_e3_3e-5/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a1e8f6e2e24b1c94242d5103eb37f00bd994202ee427d60e9b14ab8376dc0eb0
+size 16389

118_128_e3_3e-5/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b3cedabf8698995eb95156f3ff4a44dd183036ddf549201489afc7257973d1d
+size 16389

118_128_e3_3e-5/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:923f99d9b2c0025443452d5824852c01adbc19c0d815d95b351520331698a32d
+size 16389

118_128_e3_3e-5/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a562384f687ce05f3967fe7d0894d7adb6ad8c7bd20be31a7cd9eee33cb1096
+size 16389

118_128_e3_3e-5/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8030e1d75e3e1aea5ea6069299d6cd7961306cf09bbdb321484d675dfbdf7b29
+size 16389

118_128_e3_3e-5/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b1bad4f1cfbb48e3109f06b8214652e709dce1c79d7cc4433ff305452be4078
+size 1401

118_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
+  }
+}

118_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
+}

118_128_e3_3e-5/trainer_state.json

@@ -0,0 +1,853 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 585,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.02570694087403599,
+      "grad_norm": 0.698792040348053,
+      "learning_rate": 4e-06,
+      "loss": 1.4415,
+      "step": 5
+    },
+    {
+      "epoch": 0.05141388174807198,
+      "grad_norm": 0.8071721196174622,
+      "learning_rate": 9e-06,
+      "loss": 1.3829,
+      "step": 10
+    },
+    {
+      "epoch": 0.07712082262210797,
+      "grad_norm": 0.8055391907691956,
+      "learning_rate": 1.4e-05,
+      "loss": 1.4377,
+      "step": 15
+    },
+    {
+      "epoch": 0.10282776349614396,
+      "grad_norm": 0.580613374710083,
+      "learning_rate": 1.9e-05,
+      "loss": 1.3876,
+      "step": 20
+    },
+    {
+      "epoch": 0.12853470437017994,
+      "grad_norm": 0.5970566272735596,
+      "learning_rate": 2.4e-05,
+      "loss": 1.3325,
+      "step": 25
+    },
+    {
+      "epoch": 0.15424164524421594,
+      "grad_norm": 0.615569531917572,
+      "learning_rate": 2.9e-05,
+      "loss": 1.3535,
+      "step": 30
+    },
+    {
+      "epoch": 0.17994858611825193,
+      "grad_norm": 0.5293927192687988,
+      "learning_rate": 2.999615517594034e-05,
+      "loss": 1.2942,
+      "step": 35
+    },
+    {
+      "epoch": 0.20565552699228792,
+      "grad_norm": 0.5112528204917908,
+      "learning_rate": 2.998053895620817e-05,
+      "loss": 1.2591,
+      "step": 40
+    },
+    {
+      "epoch": 0.23136246786632392,
+      "grad_norm": 0.6094427108764648,
+      "learning_rate": 2.995292353804467e-05,
+      "loss": 1.2323,
+      "step": 45
+    },
+    {
+      "epoch": 0.2570694087403599,
+      "grad_norm": 0.6060751080513,
+      "learning_rate": 2.991333104100664e-05,
+      "loss": 1.1274,
+      "step": 50
+    },
+    {
+      "epoch": 0.2827763496143959,
+      "grad_norm": 0.6164885759353638,
+      "learning_rate": 2.9861793178120533e-05,
+      "loss": 1.1405,
+      "step": 55
+    },
+    {
+      "epoch": 0.30848329048843187,
+      "grad_norm": 0.7643070220947266,
+      "learning_rate": 2.9798351230480797e-05,
+      "loss": 1.051,
+      "step": 60
+    },
+    {
+      "epoch": 0.3341902313624679,
+      "grad_norm": 0.7023025155067444,
+      "learning_rate": 2.9723056014184295e-05,
+      "loss": 1.0882,
+      "step": 65
+    },
+    {
+      "epoch": 0.35989717223650386,
+      "grad_norm": 0.8972949981689453,
+      "learning_rate": 2.963596783962733e-05,
+      "loss": 1.0195,
+      "step": 70
+    },
+    {
+      "epoch": 0.3856041131105398,
+      "grad_norm": 0.8526336550712585,
+      "learning_rate": 2.953715646319787e-05,
+      "loss": 0.9768,
+      "step": 75
+    },
+    {
+      "epoch": 0.41131105398457585,
+      "grad_norm": 0.7884606719017029,
+      "learning_rate": 2.9426701031401707e-05,
+      "loss": 0.9413,
+      "step": 80
+    },
+    {
+      "epoch": 0.4370179948586118,
+      "grad_norm": 0.8249444365501404,
+      "learning_rate": 2.930469001746718e-05,
+      "loss": 0.8499,
+      "step": 85
+    },
+    {
+      "epoch": 0.46272493573264784,
+      "grad_norm": 0.9816693067550659,
+      "learning_rate": 2.9171221150479435e-05,
+      "loss": 0.853,
+      "step": 90
+    },
+    {
+      "epoch": 0.4884318766066838,
+      "grad_norm": 0.9814128875732422,
+      "learning_rate": 2.9026401337100773e-05,
+      "loss": 0.8386,
+      "step": 95
+    },
+    {
+      "epoch": 0.5141388174807198,
+      "grad_norm": 0.952884316444397,
+      "learning_rate": 2.887034657593997e-05,
+      "loss": 0.7275,
+      "step": 100
+    },
+    {
+      "epoch": 0.5398457583547558,
+      "grad_norm": 1.0664143562316895,
+      "learning_rate": 2.8703181864639013e-05,
+      "loss": 0.7596,
+      "step": 105
+    },
+    {
+      "epoch": 0.5655526992287918,
+      "grad_norm": 0.9792747497558594,
+      "learning_rate": 2.852504109975179e-05,
+      "loss": 0.702,
+      "step": 110
+    },
+    {
+      "epoch": 0.5912596401028277,
+      "grad_norm": 1.139298677444458,
+      "learning_rate": 2.8336066969494855e-05,
+      "loss": 0.7067,
+      "step": 115
+    },
+    {
+      "epoch": 0.6169665809768637,
+      "grad_norm": 1.2611732482910156,
+      "learning_rate": 2.813641083945618e-05,
+      "loss": 0.6941,
+      "step": 120
+    },
+    {
+      "epoch": 0.6426735218508998,
+      "grad_norm": 1.1369810104370117,
+      "learning_rate": 2.79262326313535e-05,
+      "loss": 0.6062,
+      "step": 125
+    },
+    {
+      "epoch": 0.6683804627249358,
+      "grad_norm": 1.1662830114364624,
+      "learning_rate": 2.7705700694939265e-05,
+      "loss": 0.6,
+      "step": 130
+    },
+    {
+      "epoch": 0.6940874035989717,
+      "grad_norm": 1.2556132078170776,
+      "learning_rate": 2.747499167315491e-05,
+      "loss": 0.5922,
+      "step": 135
+    },
+    {
+      "epoch": 0.7197943444730077,
+      "grad_norm": 1.2124667167663574,
+      "learning_rate": 2.723429036064237e-05,
+      "loss": 0.5719,
+      "step": 140
+    },
+    {
+      "epoch": 0.7455012853470437,
+      "grad_norm": 1.1445759534835815,
+      "learning_rate": 2.6983789555726178e-05,
+      "loss": 0.6012,
+      "step": 145
+    },
+    {
+      "epoch": 0.7712082262210797,
+      "grad_norm": 1.398273229598999,
+      "learning_rate": 2.6723689905984783e-05,
+      "loss": 0.5728,
+      "step": 150
+    },
+    {
+      "epoch": 0.7969151670951157,
+      "grad_norm": 1.2845197916030884,
+      "learning_rate": 2.645419974753469e-05,
+      "loss": 0.5322,
+      "step": 155
+    },
+    {
+      "epoch": 0.8226221079691517,
+      "grad_norm": 2.8561553955078125,
+      "learning_rate": 2.6175534938156176e-05,
+      "loss": 0.5,
+      "step": 160
+    },
+    {
+      "epoch": 0.8483290488431876,
+      "grad_norm": 1.3225529193878174,
+      "learning_rate": 2.5887918684394306e-05,
+      "loss": 0.4673,
+      "step": 165
+    },
+    {
+      "epoch": 0.8740359897172236,
+      "grad_norm": 1.3693724870681763,
+      "learning_rate": 2.559158136277365e-05,
+      "loss": 0.4994,
+      "step": 170
+    },
+    {
+      "epoch": 0.8997429305912596,
+      "grad_norm": 1.4422277212142944,
+      "learning_rate": 2.5286760335269973e-05,
+      "loss": 0.4783,
+      "step": 175
+    },
+    {
+      "epoch": 0.9254498714652957,
+      "grad_norm": 1.6030553579330444,
+      "learning_rate": 2.497369975918669e-05,
+      "loss": 0.4521,
+      "step": 180
+    },
+    {
+      "epoch": 0.9511568123393316,
+      "grad_norm": 1.3444982767105103,
+      "learning_rate": 2.4652650391588297e-05,
+      "loss": 0.4127,
+      "step": 185
+    },
+    {
+      "epoch": 0.9768637532133676,
+      "grad_norm": 1.2589261531829834,
+      "learning_rate": 2.4323869388447583e-05,
+      "loss": 0.4027,
+      "step": 190
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 1.9401267766952515,
+      "learning_rate": 2.3987620098667326e-05,
+      "loss": 0.4404,
+      "step": 195
+    },
+    {
+      "epoch": 1.025706940874036,
+      "grad_norm": 1.448899745941162,
+      "learning_rate": 2.364417185314158e-05,
+      "loss": 0.3748,
+      "step": 200
+    },
+    {
+      "epoch": 1.051413881748072,
+      "grad_norm": 1.2913111448287964,
+      "learning_rate": 2.329379974902545e-05,
+      "loss": 0.3163,
+      "step": 205
+    },
+    {
+      "epoch": 1.077120822622108,
+      "grad_norm": 1.667248010635376,
+      "learning_rate": 2.293678442938623e-05,
+      "loss": 0.3221,
+      "step": 210
+    },
+    {
+      "epoch": 1.1028277634961439,
+      "grad_norm": 1.2648524045944214,
+      "learning_rate": 2.2573411858412235e-05,
+      "loss": 0.3491,
+      "step": 215
+    },
+    {
+      "epoch": 1.1285347043701799,
+      "grad_norm": 1.3089914321899414,
+      "learning_rate": 2.2203973092359614e-05,
+      "loss": 0.2979,
+      "step": 220
+    },
+    {
+      "epoch": 1.154241645244216,
+      "grad_norm": 1.488169550895691,
+      "learning_rate": 2.182876404642041e-05,
+      "loss": 0.3155,
+      "step": 225
+    },
+    {
+      "epoch": 1.179948586118252,
+      "grad_norm": 1.3417556285858154,
+      "learning_rate": 2.1448085257698696e-05,
+      "loss": 0.2888,
+      "step": 230
+    },
+    {
+      "epoch": 1.205655526992288,
+      "grad_norm": 1.331659197807312,
+      "learning_rate": 2.106224164448466e-05,
+      "loss": 0.2774,
+      "step": 235
+    },
+    {
+      "epoch": 1.231362467866324,
+      "grad_norm": 1.3862390518188477,
+      "learning_rate": 2.0671542262019368e-05,
+      "loss": 0.2737,
+      "step": 240
+    },
+    {
+      "epoch": 1.2570694087403598,
+      "grad_norm": 1.2808398008346558,
+      "learning_rate": 2.0276300054945917e-05,
+      "loss": 0.2602,
+      "step": 245
+    },
+    {
+      "epoch": 1.2827763496143958,
+      "grad_norm": 1.2151902914047241,
+      "learning_rate": 1.987683160664525e-05,
+      "loss": 0.238,
+      "step": 250
+    },
+    {
+      "epoch": 1.3084832904884318,
+      "grad_norm": 1.1603834629058838,
+      "learning_rate": 1.947345688565738e-05,
+      "loss": 0.2636,
+      "step": 255
+    },
+    {
+      "epoch": 1.3341902313624678,
+      "grad_norm": 1.2844356298446655,
+      "learning_rate": 1.9066498989391107e-05,
+      "loss": 0.2613,
+      "step": 260
+    },
+    {
+      "epoch": 1.3598971722365039,
+      "grad_norm": 1.295124888420105,
+      "learning_rate": 1.8656283885327625e-05,
+      "loss": 0.2583,
+      "step": 265
+    },
+    {
+      "epoch": 1.3856041131105399,
+      "grad_norm": 1.2763663530349731,
+      "learning_rate": 1.8243140149925205e-05,
+      "loss": 0.2478,
+      "step": 270
+    },
+    {
+      "epoch": 1.411311053984576,
+      "grad_norm": 1.578310489654541,
+      "learning_rate": 1.7827398705434123e-05,
+      "loss": 0.2359,
+      "step": 275
+    },
+    {
+      "epoch": 1.437017994858612,
+      "grad_norm": 1.2663363218307495,
+      "learning_rate": 1.7409392554832657e-05,
+      "loss": 0.241,
+      "step": 280
+    },
+    {
+      "epoch": 1.462724935732648,
+      "grad_norm": 1.2047301530838013,
+      "learning_rate": 1.6989456515096436e-05,
+      "loss": 0.2298,
+      "step": 285
+    },
+    {
+      "epoch": 1.4884318766066837,
+      "grad_norm": 1.2310377359390259,
+      "learning_rate": 1.6567926949014805e-05,
+      "loss": 0.2112,
+      "step": 290
+    },
+    {
+      "epoch": 1.5141388174807198,
+      "grad_norm": 1.4305293560028076,
+      "learning_rate": 1.6145141495769022e-05,
+      "loss": 0.2203,
+      "step": 295
+    },
+    {
+      "epoch": 1.5398457583547558,
+      "grad_norm": 1.3229445219039917,
+      "learning_rate": 1.572143880048811e-05,
+      "loss": 0.2029,
+      "step": 300
+    },
+    {
+      "epoch": 1.5655526992287918,
+      "grad_norm": 1.5050373077392578,
+      "learning_rate": 1.5297158242998867e-05,
+      "loss": 0.2143,
+      "step": 305
+    },
+    {
+      "epoch": 1.5912596401028276,
+      "grad_norm": 1.3925338983535767,
+      "learning_rate": 1.487263966598752e-05,
+      "loss": 0.2008,
+      "step": 310
+    },
+    {
+      "epoch": 1.6169665809768636,
+      "grad_norm": 1.3009788990020752,
+      "learning_rate": 1.444822310279053e-05,
+      "loss": 0.2245,
+      "step": 315
+    },
+    {
+      "epoch": 1.6426735218508997,
+      "grad_norm": 1.0376914739608765,
+      "learning_rate": 1.4024248505032754e-05,
+      "loss": 0.1857,
+      "step": 320
+    },
+    {
+      "epoch": 1.6683804627249357,
+      "grad_norm": 1.3913062810897827,
+      "learning_rate": 1.3601055470331024e-05,
+      "loss": 0.1975,
+      "step": 325
+    },
+    {
+      "epoch": 1.6940874035989717,
+      "grad_norm": 1.266106128692627,
+      "learning_rate": 1.317898297028128e-05,
+      "loss": 0.1709,
+      "step": 330
+    },
+    {
+      "epoch": 1.7197943444730077,
+      "grad_norm": 1.3440873622894287,
+      "learning_rate": 1.2758369078947182e-05,
+      "loss": 0.1674,
+      "step": 335
+    },
+    {
+      "epoch": 1.7455012853470437,
+      "grad_norm": 1.1505476236343384,
+      "learning_rate": 1.2339550702067539e-05,
+      "loss": 0.1696,
+      "step": 340
+    },
+    {
+      "epoch": 1.7712082262210798,
+      "grad_norm": 1.1686441898345947,
+      "learning_rate": 1.1922863307199673e-05,
+      "loss": 0.1741,
+      "step": 345
+    },
+    {
+      "epoch": 1.7969151670951158,
+      "grad_norm": 1.2007770538330078,
+      "learning_rate": 1.150864065501464e-05,
+      "loss": 0.1624,
+      "step": 350
+    },
+    {
+      "epoch": 1.8226221079691518,
+      "grad_norm": 1.1038089990615845,
+      "learning_rate": 1.1097214531959706e-05,
+      "loss": 0.135,
+      "step": 355
+    },
+    {
+      "epoch": 1.8483290488431876,
+      "grad_norm": 1.1818947792053223,
+      "learning_rate": 1.0688914484502106e-05,
+      "loss": 0.1438,
+      "step": 360
+    },
+    {
+      "epoch": 1.8740359897172236,
+      "grad_norm": 1.2317180633544922,
+      "learning_rate": 1.0284067555167037e-05,
+      "loss": 0.1278,
+      "step": 365
+    },
+    {
+      "epoch": 1.8997429305912596,
+      "grad_norm": 1.047784447669983,
+      "learning_rate": 9.882998020581228e-06,
+      "loss": 0.1433,
+      "step": 370
+    },
+    {
+      "epoch": 1.9254498714652957,
+      "grad_norm": 1.147558331489563,
+      "learning_rate": 9.486027131732003e-06,
+      "loss": 0.123,
+      "step": 375
+    },
+    {
+      "epoch": 1.9511568123393315,
+      "grad_norm": 1.3140798807144165,
+      "learning_rate": 9.093472856649794e-06,
+      "loss": 0.1384,
+      "step": 380
+    },
+    {
+      "epoch": 1.9768637532133675,
+      "grad_norm": 1.0773415565490723,
+      "learning_rate": 8.70564962572029e-06,
+      "loss": 0.1373,
+      "step": 385
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.627793312072754,
+      "learning_rate": 8.322868079830162e-06,
+      "loss": 0.1093,
+      "step": 390
+    },
+    {
+      "epoch": 2.025706940874036,
+      "grad_norm": 0.8627418279647827,
+      "learning_rate": 7.945434821548152e-06,
+      "loss": 0.1053,
+      "step": 395
+    },
+    {
+      "epoch": 2.051413881748072,
+      "grad_norm": 0.8667323589324951,
+      "learning_rate": 7.573652169540747e-06,
+      "loss": 0.1003,
+      "step": 400
+    },
+    {
+      "epoch": 2.077120822622108,
+      "grad_norm": 1.0445892810821533,
+      "learning_rate": 7.207817916419276e-06,
+      "loss": 0.1067,
+      "step": 405
+    },
+    {
+      "epoch": 2.102827763496144,
+      "grad_norm": 1.1511319875717163,
+      "learning_rate": 6.848225090212234e-06,
+      "loss": 0.0911,
+      "step": 410
+    },
+    {
+      "epoch": 2.12853470437018,
+      "grad_norm": 0.9618096351623535,
+      "learning_rate": 6.4951617196540244e-06,
+      "loss": 0.0862,
+      "step": 415
+    },
+    {
+      "epoch": 2.154241645244216,
+      "grad_norm": 1.1876215934753418,
+      "learning_rate": 6.148910603478027e-06,
+      "loss": 0.1043,
+      "step": 420
+    },
+    {
+      "epoch": 2.179948586118252,
+      "grad_norm": 0.7668461799621582,
+      "learning_rate": 5.8097490838988365e-06,
+      "loss": 0.0908,
+      "step": 425
+    },
+    {
+      "epoch": 2.2056555269922877,
+      "grad_norm": 0.9387179613113403,
+      "learning_rate": 5.477948824465118e-06,
+      "loss": 0.0904,
+      "step": 430
+    },
+    {
+      "epoch": 2.2313624678663238,
+      "grad_norm": 0.8298426270484924,
+      "learning_rate": 5.153775592460958e-06,
+      "loss": 0.0854,
+      "step": 435
+    },
+    {
+      "epoch": 2.2570694087403598,
+      "grad_norm": 0.8625733256340027,
+      "learning_rate": 4.837489046030053e-06,
+      "loss": 0.0911,
+      "step": 440
+    },
+    {
+      "epoch": 2.282776349614396,
+      "grad_norm": 1.2923249006271362,
+      "learning_rate": 4.529342526193235e-06,
+      "loss": 0.0921,
+      "step": 445
+    },
+    {
+      "epoch": 2.308483290488432,
+      "grad_norm": 0.9703097939491272,
+      "learning_rate": 4.229582853925924e-06,
+      "loss": 0.0955,
+      "step": 450
+    },
+    {
+      "epoch": 2.334190231362468,
+      "grad_norm": 1.0343331098556519,
+      "learning_rate": 3.938450132458028e-06,
+      "loss": 0.0818,
+      "step": 455
+    },
+    {
+      "epoch": 2.359897172236504,
+      "grad_norm": 0.7084493637084961,
+      "learning_rate": 3.656177554954693e-06,
+      "loss": 0.0795,
+      "step": 460
+    },
+    {
+      "epoch": 2.38560411311054,
+      "grad_norm": 0.9468570947647095,
+      "learning_rate": 3.3829912177318863e-06,
+      "loss": 0.0739,
+      "step": 465
+    },
+    {
+      "epoch": 2.411311053984576,
+      "grad_norm": 0.8550546765327454,
+      "learning_rate": 3.1191099391564665e-06,
+      "loss": 0.0791,
+      "step": 470
+    },
+    {
+      "epoch": 2.437017994858612,
+      "grad_norm": 0.9636321663856506,
+      "learning_rate": 2.86474508437579e-06,
+      "loss": 0.0771,
+      "step": 475
+    },
+    {
+      "epoch": 2.462724935732648,
+      "grad_norm": 0.9645887613296509,
+      "learning_rate": 2.6201003960172327e-06,
+      "loss": 0.0738,
+      "step": 480
+    },
+    {
+      "epoch": 2.488431876606684,
+      "grad_norm": 0.8179371953010559,
+      "learning_rate": 2.385371830993251e-06,
+      "loss": 0.0825,
+      "step": 485
+    },
+    {
+      "epoch": 2.5141388174807195,
+      "grad_norm": 0.8026527762413025,
+      "learning_rate": 2.160747403542686e-06,
+      "loss": 0.0695,
+      "step": 490
+    },
+    {
+      "epoch": 2.5398457583547556,
+      "grad_norm": 0.7437753081321716,
+      "learning_rate": 1.9464070346340203e-06,
+      "loss": 0.0633,
+      "step": 495
+    },
+    {
+      "epoch": 2.5655526992287916,
+      "grad_norm": 0.8937798142433167,
+      "learning_rate": 1.7425224078512674e-06,
+      "loss": 0.0834,
+      "step": 500
+    },
+    {
+      "epoch": 2.5912596401028276,
+      "grad_norm": 0.6792085766792297,
+      "learning_rate": 1.549256831877846e-06,
+      "loss": 0.0751,
+      "step": 505
+    },
+    {
+      "epoch": 2.6169665809768636,
+      "grad_norm": 0.9352570176124573,
+      "learning_rate": 1.366765109688658e-06,
+      "loss": 0.0813,
+      "step": 510
+    },
+    {
+      "epoch": 2.6426735218508997,
+      "grad_norm": 0.6518654227256775,
+      "learning_rate": 1.1951934145551086e-06,
+      "loss": 0.0632,
+      "step": 515
+    },
+    {
+      "epoch": 2.6683804627249357,
+      "grad_norm": 0.8614311218261719,
+      "learning_rate": 1.0346791729623928e-06,
+      "loss": 0.0752,
+      "step": 520
+    },
+    {
+      "epoch": 2.6940874035989717,
+      "grad_norm": 0.902529239654541,
+      "learning_rate": 8.853509545328442e-07,
+      "loss": 0.072,
+      "step": 525
+    },
+    {
+      "epoch": 2.7197943444730077,
+      "grad_norm": 0.8625785112380981,
+      "learning_rate": 7.473283690434907e-07,
+      "loss": 0.0717,
+      "step": 530
+    },
+    {
+      "epoch": 2.7455012853470437,
+      "grad_norm": 0.7356880903244019,
+      "learning_rate": 6.207219706203326e-07,
+      "loss": 0.0711,
+      "step": 535
+    },
+    {
+      "epoch": 2.7712082262210798,
+      "grad_norm": 0.8897441029548645,
+      "learning_rate": 5.056331691860577e-07,
+      "loss": 0.0782,
+      "step": 540
+    },
+    {
+      "epoch": 2.796915167095116,
+      "grad_norm": 0.6261716485023499,
+      "learning_rate": 4.0215414923213287e-07,
+      "loss": 0.0588,
+      "step": 545
+    },
+    {
+      "epoch": 2.822622107969152,
+      "grad_norm": 0.7765554785728455,
+      "learning_rate": 3.1036779598034914e-07,
+      "loss": 0.0679,
+      "step": 550
+    },
+    {
+      "epoch": 2.848329048843188,
+      "grad_norm": 0.7407102584838867,
+      "learning_rate": 2.3034762899293294e-07,
+      "loss": 0.0717,
+      "step": 555
+    },
+    {
+      "epoch": 2.874035989717224,
+      "grad_norm": 0.8059044480323792,
+      "learning_rate": 1.6215774328442012e-07,
+      "loss": 0.0832,
+      "step": 560
+    },
+    {
+      "epoch": 2.89974293059126,
+      "grad_norm": 0.7390157580375671,
+      "learning_rate": 1.0585275798246974e-07,
+      "loss": 0.0688,
+      "step": 565
+    },
+    {
+      "epoch": 2.925449871465296,
+      "grad_norm": 0.8127499222755432,
+      "learning_rate": 6.147777257871468e-08,
+      "loss": 0.0733,
+      "step": 570
+    },
+    {
+      "epoch": 2.9511568123393315,
+      "grad_norm": 0.9213354587554932,
+      "learning_rate": 2.9068330804711276e-08,
+      "loss": 0.0757,
+      "step": 575
+    },
+    {
+      "epoch": 2.9768637532133675,
+      "grad_norm": 0.7793170809745789,
+      "learning_rate": 8.650392161921317e-09,
+      "loss": 0.0751,
+      "step": 580
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.9651294350624084,
+      "learning_rate": 2.403111285115278e-10,
+      "loss": 0.0763,
+      "step": 585
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 585,
+  "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": 7.680557859293102e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

118_128_e3_3e-5/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ae7c0ee0948b060b85307f31ac3914d995dd54e3073f4ac31564c4b32cf109c
+size 8209

118_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)