Upload folder using huggingface_hub

114_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",
+    "v_proj",
+    "up_proj",
+    "down_proj",
+    "gate_proj",
+    "o_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

114_128_e3_3e-5/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84347a60c0822d08dc39f53d369530a9f36da2d4c826205f6fb7e580de9c93ac
+size 791751704

114_128_e3_3e-5/latest

@@ -0,0 +1 @@
+global_step642

114_128_e3_3e-5/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b885fdc8863c57dd1006763e878d835803058c16fd468cc2e1d28c0bd635df15
+size 16389

114_128_e3_3e-5/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b07b937d8c4366628c23d4ffac5160ad156e2f116a2bb1d4a3f441b565fba22
+size 16389

114_128_e3_3e-5/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6ff7be3af8d77f8ff4627ef0454a8eb6f54d09732d12b63670c0a6edbf37662
+size 16389

114_128_e3_3e-5/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70086267dba3d2aa6bd87f9705d99232803aae2212d8177ca56fe35fbc64e9f3
+size 16389

114_128_e3_3e-5/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f107e602c5d5c3777d94b1e53b3be3ea82bad047a862cd8e2558d28838b31095
+size 16389

114_128_e3_3e-5/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca31df6da5c767e4e7d2acf4e6aa800f8426a23e9914366411e9beb89cf7375b
+size 16389

114_128_e3_3e-5/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:326bab8ccda58eb26ae4c222758ad59abe675b1242c30d8989e0449f018bb394
+size 16389

114_128_e3_3e-5/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5face2f963f5efa9675432fa16c415d8dd2583d8393a4d6d2b513fdd254521d7
+size 16389

114_128_e3_3e-5/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f2a4d841daa425ea41b014f79f5137749cb19964ebe6a7408e345fa674768c8b
+size 1401

114_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
+  }
+}

114_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
+}

114_128_e3_3e-5/trainer_state.json

@@ -0,0 +1,930 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 642,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.02336448598130841,
+      "grad_norm": 0.9358441233634949,
+      "learning_rate": 3.6363636363636366e-06,
+      "loss": 1.3716,
+      "step": 5
+    },
+    {
+      "epoch": 0.04672897196261682,
+      "grad_norm": 0.786108672618866,
+      "learning_rate": 8.181818181818181e-06,
+      "loss": 1.3674,
+      "step": 10
+    },
+    {
+      "epoch": 0.07009345794392523,
+      "grad_norm": 0.7749671936035156,
+      "learning_rate": 1.2727272727272728e-05,
+      "loss": 1.3376,
+      "step": 15
+    },
+    {
+      "epoch": 0.09345794392523364,
+      "grad_norm": 0.5120208859443665,
+      "learning_rate": 1.7272727272727274e-05,
+      "loss": 1.3301,
+      "step": 20
+    },
+    {
+      "epoch": 0.11682242990654206,
+      "grad_norm": 0.6779114603996277,
+      "learning_rate": 2.1818181818181818e-05,
+      "loss": 1.2928,
+      "step": 25
+    },
+    {
+      "epoch": 0.14018691588785046,
+      "grad_norm": 0.5924385786056519,
+      "learning_rate": 2.6363636363636365e-05,
+      "loss": 1.2035,
+      "step": 30
+    },
+    {
+      "epoch": 0.16355140186915887,
+      "grad_norm": 0.6084502339363098,
+      "learning_rate": 2.9999800416120386e-05,
+      "loss": 1.3186,
+      "step": 35
+    },
+    {
+      "epoch": 0.18691588785046728,
+      "grad_norm": 0.6312307119369507,
+      "learning_rate": 2.9992815537990055e-05,
+      "loss": 1.1989,
+      "step": 40
+    },
+    {
+      "epoch": 0.2102803738317757,
+      "grad_norm": 0.5338754653930664,
+      "learning_rate": 2.9975856776406864e-05,
+      "loss": 1.1806,
+      "step": 45
+    },
+    {
+      "epoch": 0.2336448598130841,
+      "grad_norm": 0.6062021255493164,
+      "learning_rate": 2.9948935413088454e-05,
+      "loss": 1.1187,
+      "step": 50
+    },
+    {
+      "epoch": 0.2570093457943925,
+      "grad_norm": 0.8071431517601013,
+      "learning_rate": 2.9912069357315394e-05,
+      "loss": 1.0957,
+      "step": 55
+    },
+    {
+      "epoch": 0.2803738317757009,
+      "grad_norm": 0.621833324432373,
+      "learning_rate": 2.9865283134017167e-05,
+      "loss": 1.1689,
+      "step": 60
+    },
+    {
+      "epoch": 0.3037383177570093,
+      "grad_norm": 0.7238261699676514,
+      "learning_rate": 2.980860786745706e-05,
+      "loss": 1.0123,
+      "step": 65
+    },
+    {
+      "epoch": 0.32710280373831774,
+      "grad_norm": 0.6746344566345215,
+      "learning_rate": 2.9742081260526968e-05,
+      "loss": 1.0454,
+      "step": 70
+    },
+    {
+      "epoch": 0.35046728971962615,
+      "grad_norm": 1.0194181203842163,
+      "learning_rate": 2.9665747569665735e-05,
+      "loss": 0.9863,
+      "step": 75
+    },
+    {
+      "epoch": 0.37383177570093457,
+      "grad_norm": 0.9060932993888855,
+      "learning_rate": 2.957965757541784e-05,
+      "loss": 0.9886,
+      "step": 80
+    },
+    {
+      "epoch": 0.397196261682243,
+      "grad_norm": 0.8628988265991211,
+      "learning_rate": 2.948386854865191e-05,
+      "loss": 0.9234,
+      "step": 85
+    },
+    {
+      "epoch": 0.4205607476635514,
+      "grad_norm": 0.7447900772094727,
+      "learning_rate": 2.937844421246162e-05,
+      "loss": 0.8827,
+      "step": 90
+    },
+    {
+      "epoch": 0.4439252336448598,
+      "grad_norm": 0.9761838316917419,
+      "learning_rate": 2.926345469977428e-05,
+      "loss": 0.8834,
+      "step": 95
+    },
+    {
+      "epoch": 0.4672897196261682,
+      "grad_norm": 1.2164639234542847,
+      "learning_rate": 2.9138976506695307e-05,
+      "loss": 0.8209,
+      "step": 100
+    },
+    {
+      "epoch": 0.49065420560747663,
+      "grad_norm": 0.8026975393295288,
+      "learning_rate": 2.9005092441619627e-05,
+      "loss": 0.8797,
+      "step": 105
+    },
+    {
+      "epoch": 0.514018691588785,
+      "grad_norm": 1.0448005199432373,
+      "learning_rate": 2.8861891570143885e-05,
+      "loss": 0.8235,
+      "step": 110
+    },
+    {
+      "epoch": 0.5373831775700935,
+      "grad_norm": 0.9038024544715881,
+      "learning_rate": 2.8709469155816066e-05,
+      "loss": 0.7449,
+      "step": 115
+    },
+    {
+      "epoch": 0.5607476635514018,
+      "grad_norm": 0.9512847065925598,
+      "learning_rate": 2.854792659676198e-05,
+      "loss": 0.716,
+      "step": 120
+    },
+    {
+      "epoch": 0.5841121495327103,
+      "grad_norm": 1.101916790008545,
+      "learning_rate": 2.8377371358230733e-05,
+      "loss": 0.7176,
+      "step": 125
+    },
+    {
+      "epoch": 0.6074766355140186,
+      "grad_norm": 1.088397741317749,
+      "learning_rate": 2.819791690110412e-05,
+      "loss": 0.6956,
+      "step": 130
+    },
+    {
+      "epoch": 0.6308411214953271,
+      "grad_norm": 1.0392569303512573,
+      "learning_rate": 2.8009682606417397e-05,
+      "loss": 0.7005,
+      "step": 135
+    },
+    {
+      "epoch": 0.6542056074766355,
+      "grad_norm": 1.1375138759613037,
+      "learning_rate": 2.781279369594177e-05,
+      "loss": 0.627,
+      "step": 140
+    },
+    {
+      "epoch": 0.677570093457944,
+      "grad_norm": 1.361741065979004,
+      "learning_rate": 2.7607381148881345e-05,
+      "loss": 0.5745,
+      "step": 145
+    },
+    {
+      "epoch": 0.7009345794392523,
+      "grad_norm": 1.3444445133209229,
+      "learning_rate": 2.7393581614739924e-05,
+      "loss": 0.5811,
+      "step": 150
+    },
+    {
+      "epoch": 0.7242990654205608,
+      "grad_norm": 1.291254997253418,
+      "learning_rate": 2.7171537322415777e-05,
+      "loss": 0.5472,
+      "step": 155
+    },
+    {
+      "epoch": 0.7476635514018691,
+      "grad_norm": 1.3581788539886475,
+      "learning_rate": 2.6941395985584656e-05,
+      "loss": 0.5723,
+      "step": 160
+    },
+    {
+      "epoch": 0.7710280373831776,
+      "grad_norm": 1.1932355165481567,
+      "learning_rate": 2.6703310704434144e-05,
+      "loss": 0.5871,
+      "step": 165
+    },
+    {
+      "epoch": 0.794392523364486,
+      "grad_norm": 1.4491585493087769,
+      "learning_rate": 2.6457439863814637e-05,
+      "loss": 0.5326,
+      "step": 170
+    },
+    {
+      "epoch": 0.8177570093457944,
+      "grad_norm": 1.2222353219985962,
+      "learning_rate": 2.6203947027874753e-05,
+      "loss": 0.5424,
+      "step": 175
+    },
+    {
+      "epoch": 0.8411214953271028,
+      "grad_norm": 1.4209558963775635,
+      "learning_rate": 2.5943000831251215e-05,
+      "loss": 0.5247,
+      "step": 180
+    },
+    {
+      "epoch": 0.8644859813084113,
+      "grad_norm": 1.2883238792419434,
+      "learning_rate": 2.5674774866885664e-05,
+      "loss": 0.464,
+      "step": 185
+    },
+    {
+      "epoch": 0.8878504672897196,
+      "grad_norm": 1.1829696893692017,
+      "learning_rate": 2.539944757054293e-05,
+      "loss": 0.478,
+      "step": 190
+    },
+    {
+      "epoch": 0.9112149532710281,
+      "grad_norm": 1.3029143810272217,
+      "learning_rate": 2.5117202102107707e-05,
+      "loss": 0.4824,
+      "step": 195
+    },
+    {
+      "epoch": 0.9345794392523364,
+      "grad_norm": 1.1535396575927734,
+      "learning_rate": 2.4828226223738493e-05,
+      "loss": 0.4105,
+      "step": 200
+    },
+    {
+      "epoch": 0.9579439252336449,
+      "grad_norm": 1.3797191381454468,
+      "learning_rate": 2.4532712174959882e-05,
+      "loss": 0.438,
+      "step": 205
+    },
+    {
+      "epoch": 0.9813084112149533,
+      "grad_norm": 1.3636231422424316,
+      "learning_rate": 2.4230856544776377e-05,
+      "loss": 0.4146,
+      "step": 210
+    },
+    {
+      "epoch": 1.0046728971962617,
+      "grad_norm": 1.3273917436599731,
+      "learning_rate": 2.3922860140892667e-05,
+      "loss": 0.3744,
+      "step": 215
+    },
+    {
+      "epoch": 1.02803738317757,
+      "grad_norm": 1.480552315711975,
+      "learning_rate": 2.3608927856127483e-05,
+      "loss": 0.3295,
+      "step": 220
+    },
+    {
+      "epoch": 1.0514018691588785,
+      "grad_norm": 1.2728606462478638,
+      "learning_rate": 2.3289268532109828e-05,
+      "loss": 0.3603,
+      "step": 225
+    },
+    {
+      "epoch": 1.074766355140187,
+      "grad_norm": 1.278167486190796,
+      "learning_rate": 2.2964094820348302e-05,
+      "loss": 0.3287,
+      "step": 230
+    },
+    {
+      "epoch": 1.0981308411214954,
+      "grad_norm": 1.105540156364441,
+      "learning_rate": 2.2633623040765936e-05,
+      "loss": 0.3218,
+      "step": 235
+    },
+    {
+      "epoch": 1.1214953271028036,
+      "grad_norm": 1.4142242670059204,
+      "learning_rate": 2.2298073037794585e-05,
+      "loss": 0.3223,
+      "step": 240
+    },
+    {
+      "epoch": 1.144859813084112,
+      "grad_norm": 1.3108631372451782,
+      "learning_rate": 2.1957668034124732e-05,
+      "loss": 0.3235,
+      "step": 245
+    },
+    {
+      "epoch": 1.1682242990654206,
+      "grad_norm": 1.4890906810760498,
+      "learning_rate": 2.1612634482207874e-05,
+      "loss": 0.2706,
+      "step": 250
+    },
+    {
+      "epoch": 1.191588785046729,
+      "grad_norm": 1.455453872680664,
+      "learning_rate": 2.1263201913610343e-05,
+      "loss": 0.297,
+      "step": 255
+    },
+    {
+      "epoch": 1.2149532710280373,
+      "grad_norm": 1.3076280355453491,
+      "learning_rate": 2.0909602786318757e-05,
+      "loss": 0.2757,
+      "step": 260
+    },
+    {
+      "epoch": 1.2383177570093458,
+      "grad_norm": 1.2770133018493652,
+      "learning_rate": 2.055207233009872e-05,
+      "loss": 0.2825,
+      "step": 265
+    },
+    {
+      "epoch": 1.2616822429906542,
+      "grad_norm": 1.4440333843231201,
+      "learning_rate": 2.0190848390009556e-05,
+      "loss": 0.2689,
+      "step": 270
+    },
+    {
+      "epoch": 1.2850467289719627,
+      "grad_norm": 1.281765103340149,
+      "learning_rate": 1.9826171268179302e-05,
+      "loss": 0.2522,
+      "step": 275
+    },
+    {
+      "epoch": 1.308411214953271,
+      "grad_norm": 1.4272243976593018,
+      "learning_rate": 1.9458283563945063e-05,
+      "loss": 0.2408,
+      "step": 280
+    },
+    {
+      "epoch": 1.3317757009345794,
+      "grad_norm": 1.3573485612869263,
+      "learning_rate": 1.9087430012465264e-05,
+      "loss": 0.2466,
+      "step": 285
+    },
+    {
+      "epoch": 1.355140186915888,
+      "grad_norm": 1.5090835094451904,
+      "learning_rate": 1.8713857321910968e-05,
+      "loss": 0.2668,
+      "step": 290
+    },
+    {
+      "epoch": 1.3785046728971961,
+      "grad_norm": 1.6274574995040894,
+      "learning_rate": 1.8337814009344716e-05,
+      "loss": 0.256,
+      "step": 295
+    },
+    {
+      "epoch": 1.4018691588785046,
+      "grad_norm": 1.3695251941680908,
+      "learning_rate": 1.7959550235396002e-05,
+      "loss": 0.2412,
+      "step": 300
+    },
+    {
+      "epoch": 1.425233644859813,
+      "grad_norm": 1.5408960580825806,
+      "learning_rate": 1.7579317637843344e-05,
+      "loss": 0.2352,
+      "step": 305
+    },
+    {
+      "epoch": 1.4485981308411215,
+      "grad_norm": 1.0863654613494873,
+      "learning_rate": 1.719736916421374e-05,
+      "loss": 0.1999,
+      "step": 310
+    },
+    {
+      "epoch": 1.47196261682243,
+      "grad_norm": 1.133382797241211,
+      "learning_rate": 1.6813958903510767e-05,
+      "loss": 0.1938,
+      "step": 315
+    },
+    {
+      "epoch": 1.4953271028037383,
+      "grad_norm": 1.3420647382736206,
+      "learning_rate": 1.6429341917183377e-05,
+      "loss": 0.2018,
+      "step": 320
+    },
+    {
+      "epoch": 1.5186915887850467,
+      "grad_norm": 1.3628230094909668,
+      "learning_rate": 1.604377406944772e-05,
+      "loss": 0.2154,
+      "step": 325
+    },
+    {
+      "epoch": 1.542056074766355,
+      "grad_norm": 1.4580873250961304,
+      "learning_rate": 1.5657511857074988e-05,
+      "loss": 0.1887,
+      "step": 330
+    },
+    {
+      "epoch": 1.5654205607476634,
+      "grad_norm": 0.9670282006263733,
+      "learning_rate": 1.5270812238758407e-05,
+      "loss": 0.236,
+      "step": 335
+    },
+    {
+      "epoch": 1.588785046728972,
+      "grad_norm": 1.3702319860458374,
+      "learning_rate": 1.4883932464172968e-05,
+      "loss": 0.1934,
+      "step": 340
+    },
+    {
+      "epoch": 1.6121495327102804,
+      "grad_norm": 1.4708689451217651,
+      "learning_rate": 1.4497129902841576e-05,
+      "loss": 0.1879,
+      "step": 345
+    },
+    {
+      "epoch": 1.6355140186915889,
+      "grad_norm": 1.2887157201766968,
+      "learning_rate": 1.411066187292147e-05,
+      "loss": 0.1823,
+      "step": 350
+    },
+    {
+      "epoch": 1.6588785046728973,
+      "grad_norm": 1.2889783382415771,
+      "learning_rate": 1.3724785470024793e-05,
+      "loss": 0.2185,
+      "step": 355
+    },
+    {
+      "epoch": 1.6822429906542056,
+      "grad_norm": 1.0225321054458618,
+      "learning_rate": 1.333975739618727e-05,
+      "loss": 0.1747,
+      "step": 360
+    },
+    {
+      "epoch": 1.705607476635514,
+      "grad_norm": 1.0174130201339722,
+      "learning_rate": 1.2955833789098634e-05,
+      "loss": 0.1758,
+      "step": 365
+    },
+    {
+      "epoch": 1.7289719626168223,
+      "grad_norm": 1.310351848602295,
+      "learning_rate": 1.2573270051708529e-05,
+      "loss": 0.1688,
+      "step": 370
+    },
+    {
+      "epoch": 1.7523364485981308,
+      "grad_norm": 1.2477644681930542,
+      "learning_rate": 1.2192320682321189e-05,
+      "loss": 0.1674,
+      "step": 375
+    },
+    {
+      "epoch": 1.7757009345794392,
+      "grad_norm": 1.1465882062911987,
+      "learning_rate": 1.18132391052919e-05,
+      "loss": 0.1536,
+      "step": 380
+    },
+    {
+      "epoch": 1.7990654205607477,
+      "grad_norm": 1.2297980785369873,
+      "learning_rate": 1.1436277502437936e-05,
+      "loss": 0.1719,
+      "step": 385
+    },
+    {
+      "epoch": 1.8224299065420562,
+      "grad_norm": 1.106257677078247,
+      "learning_rate": 1.1061686645276063e-05,
+      "loss": 0.1647,
+      "step": 390
+    },
+    {
+      "epoch": 1.8457943925233646,
+      "grad_norm": 1.1262325048446655,
+      "learning_rate": 1.0689715728198248e-05,
+      "loss": 0.1756,
+      "step": 395
+    },
+    {
+      "epoch": 1.8691588785046729,
+      "grad_norm": 1.0365469455718994,
+      "learning_rate": 1.0320612202696533e-05,
+      "loss": 0.1316,
+      "step": 400
+    },
+    {
+      "epoch": 1.8925233644859814,
+      "grad_norm": 1.127778172492981,
+      "learning_rate": 9.954621612747371e-06,
+      "loss": 0.1332,
+      "step": 405
+    },
+    {
+      "epoch": 1.9158878504672896,
+      "grad_norm": 1.2969752550125122,
+      "learning_rate": 9.591987431464917e-06,
+      "loss": 0.1393,
+      "step": 410
+    },
+    {
+      "epoch": 1.939252336448598,
+      "grad_norm": 1.208370566368103,
+      "learning_rate": 9.232950899131958e-06,
+      "loss": 0.1579,
+      "step": 415
+    },
+    {
+      "epoch": 1.9626168224299065,
+      "grad_norm": 1.1617950201034546,
+      "learning_rate": 8.877750862716208e-06,
+      "loss": 0.1557,
+      "step": 420
+    },
+    {
+      "epoch": 1.985981308411215,
+      "grad_norm": 1.0477648973464966,
+      "learning_rate": 8.526623616978763e-06,
+      "loss": 0.1364,
+      "step": 425
+    },
+    {
+      "epoch": 2.0093457943925235,
+      "grad_norm": 0.9385583996772766,
+      "learning_rate": 8.179802747280367e-06,
+      "loss": 0.1237,
+      "step": 430
+    },
+    {
+      "epoch": 2.032710280373832,
+      "grad_norm": 1.0331815481185913,
+      "learning_rate": 7.837518974190136e-06,
+      "loss": 0.0961,
+      "step": 435
+    },
+    {
+      "epoch": 2.05607476635514,
+      "grad_norm": 1.0379167795181274,
+      "learning_rate": 7.500000000000004e-06,
+      "loss": 0.0971,
+      "step": 440
+    },
+    {
+      "epoch": 2.0794392523364484,
+      "grad_norm": 1.0979682207107544,
+      "learning_rate": 7.167470357247111e-06,
+      "loss": 0.1027,
+      "step": 445
+    },
+    {
+      "epoch": 2.102803738317757,
+      "grad_norm": 1.0050503015518188,
+      "learning_rate": 6.840151259344844e-06,
+      "loss": 0.0998,
+      "step": 450
+    },
+    {
+      "epoch": 2.1261682242990654,
+      "grad_norm": 0.9471852779388428,
+      "learning_rate": 6.5182604534218964e-06,
+      "loss": 0.0982,
+      "step": 455
+    },
+    {
+      "epoch": 2.149532710280374,
+      "grad_norm": 1.020105242729187,
+      "learning_rate": 6.202012075467255e-06,
+      "loss": 0.1072,
+      "step": 460
+    },
+    {
+      "epoch": 2.1728971962616823,
+      "grad_norm": 1.0183130502700806,
+      "learning_rate": 5.891616507877458e-06,
+      "loss": 0.1019,
+      "step": 465
+    },
+    {
+      "epoch": 2.196261682242991,
+      "grad_norm": 1.0451428890228271,
+      "learning_rate": 5.587280239500953e-06,
+      "loss": 0.0957,
+      "step": 470
+    },
+    {
+      "epoch": 2.2196261682242993,
+      "grad_norm": 1.2281867265701294,
+      "learning_rate": 5.289205728272587e-06,
+      "loss": 0.0778,
+      "step": 475
+    },
+    {
+      "epoch": 2.2429906542056073,
+      "grad_norm": 0.8628813624382019,
+      "learning_rate": 4.997591266529623e-06,
+      "loss": 0.096,
+      "step": 480
+    },
+    {
+      "epoch": 2.2663551401869158,
+      "grad_norm": 0.777028501033783,
+      "learning_rate": 4.7126308490989334e-06,
+      "loss": 0.0927,
+      "step": 485
+    },
+    {
+      "epoch": 2.289719626168224,
+      "grad_norm": 1.0281288623809814,
+      "learning_rate": 4.434514044243055e-06,
+      "loss": 0.0878,
+      "step": 490
+    },
+    {
+      "epoch": 2.3130841121495327,
+      "grad_norm": 0.7649292349815369,
+      "learning_rate": 4.16342586755102e-06,
+      "loss": 0.085,
+      "step": 495
+    },
+    {
+      "epoch": 2.336448598130841,
+      "grad_norm": 1.2111896276474,
+      "learning_rate": 3.899546658857791e-06,
+      "loss": 0.098,
+      "step": 500
+    },
+    {
+      "epoch": 2.3598130841121496,
+      "grad_norm": 0.9031320214271545,
+      "learning_rate": 3.6430519622742387e-06,
+      "loss": 0.0843,
+      "step": 505
+    },
+    {
+      "epoch": 2.383177570093458,
+      "grad_norm": 0.7608156800270081,
+      "learning_rate": 3.394112409407455e-06,
+      "loss": 0.0802,
+      "step": 510
+    },
+    {
+      "epoch": 2.406542056074766,
+      "grad_norm": 0.9243894219398499,
+      "learning_rate": 3.152893605849053e-06,
+      "loss": 0.0802,
+      "step": 515
+    },
+    {
+      "epoch": 2.4299065420560746,
+      "grad_norm": 0.7292394638061523,
+      "learning_rate": 2.9195560210070117e-06,
+      "loss": 0.0879,
+      "step": 520
+    },
+    {
+      "epoch": 2.453271028037383,
+      "grad_norm": 0.7857130765914917,
+      "learning_rate": 2.6942548813543354e-06,
+      "loss": 0.0817,
+      "step": 525
+    },
+    {
+      "epoch": 2.4766355140186915,
+      "grad_norm": 0.9687077403068542,
+      "learning_rate": 2.4771400671655413e-06,
+      "loss": 0.074,
+      "step": 530
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 0.8597396016120911,
+      "learning_rate": 2.268356012809696e-06,
+      "loss": 0.0832,
+      "step": 535
+    },
+    {
+      "epoch": 2.5233644859813085,
+      "grad_norm": 0.8148025870323181,
+      "learning_rate": 2.06804161066626e-06,
+      "loss": 0.0711,
+      "step": 540
+    },
+    {
+      "epoch": 2.546728971962617,
+      "grad_norm": 0.761853039264679,
+      "learning_rate": 1.8763301187277554e-06,
+      "loss": 0.0808,
+      "step": 545
+    },
+    {
+      "epoch": 2.5700934579439254,
+      "grad_norm": 0.8312852382659912,
+      "learning_rate": 1.69334907195067e-06,
+      "loss": 0.0874,
+      "step": 550
+    },
+    {
+      "epoch": 2.593457943925234,
+      "grad_norm": 0.7686761021614075,
+      "learning_rate": 1.5192201974135506e-06,
+      "loss": 0.0681,
+      "step": 555
+    },
+    {
+      "epoch": 2.616822429906542,
+      "grad_norm": 0.7953792214393616,
+      "learning_rate": 1.354059333338773e-06,
+      "loss": 0.0832,
+      "step": 560
+    },
+    {
+      "epoch": 2.6401869158878504,
+      "grad_norm": 0.6809560060501099,
+      "learning_rate": 1.197976352031851e-06,
+      "loss": 0.0749,
+      "step": 565
+    },
+    {
+      "epoch": 2.663551401869159,
+      "grad_norm": 0.795145571231842,
+      "learning_rate": 1.0510750867895152e-06,
+      "loss": 0.0769,
+      "step": 570
+    },
+    {
+      "epoch": 2.6869158878504673,
+      "grad_norm": 1.0139777660369873,
+      "learning_rate": 9.13453262825209e-07,
+      "loss": 0.0805,
+      "step": 575
+    },
+    {
+      "epoch": 2.710280373831776,
+      "grad_norm": 0.7160447239875793,
+      "learning_rate": 7.852024322579649e-07,
+      "loss": 0.0807,
+      "step": 580
+    },
+    {
+      "epoch": 2.7336448598130842,
+      "grad_norm": 0.6345763802528381,
+      "learning_rate": 6.664079132078881e-07,
+      "loss": 0.0777,
+      "step": 585
+    },
+    {
+      "epoch": 2.7570093457943923,
+      "grad_norm": 0.8491983413696289,
+      "learning_rate": 5.571487330387803e-07,
+      "loss": 0.0806,
+      "step": 590
+    },
+    {
+      "epoch": 2.7803738317757007,
+      "grad_norm": 0.710375189781189,
+      "learning_rate": 4.574975757856287e-07,
+      "loss": 0.075,
+      "step": 595
+    },
+    {
+      "epoch": 2.803738317757009,
+      "grad_norm": 0.706425130367279,
+      "learning_rate": 3.6752073380198314e-07,
+      "loss": 0.0719,
+      "step": 600
+    },
+    {
+      "epoch": 2.8271028037383177,
+      "grad_norm": 0.8771233558654785,
+      "learning_rate": 2.8727806365935806e-07,
+      "loss": 0.0798,
+      "step": 605
+    },
+    {
+      "epoch": 2.850467289719626,
+      "grad_norm": 0.8301739692687988,
+      "learning_rate": 2.1682294632798582e-07,
+      "loss": 0.0744,
+      "step": 610
+    },
+    {
+      "epoch": 2.8738317757009346,
+      "grad_norm": 0.8209679126739502,
+      "learning_rate": 1.5620225166544155e-07,
+      "loss": 0.0695,
+      "step": 615
+    },
+    {
+      "epoch": 2.897196261682243,
+      "grad_norm": 0.6941513419151306,
+      "learning_rate": 1.0545630723674093e-07,
+      "loss": 0.0737,
+      "step": 620
+    },
+    {
+      "epoch": 2.9205607476635516,
+      "grad_norm": 0.6955289244651794,
+      "learning_rate": 6.461887148666523e-08,
+      "loss": 0.072,
+      "step": 625
+    },
+    {
+      "epoch": 2.94392523364486,
+      "grad_norm": 0.7833700776100159,
+      "learning_rate": 3.371711128214883e-08,
+      "loss": 0.0807,
+      "step": 630
+    },
+    {
+      "epoch": 2.9672897196261685,
+      "grad_norm": 0.724115252494812,
+      "learning_rate": 1.2771583839679136e-08,
+      "loss": 0.0738,
+      "step": 635
+    },
+    {
+      "epoch": 2.9906542056074765,
+      "grad_norm": 0.7053651213645935,
+      "learning_rate": 1.7962230497275966e-09,
+      "loss": 0.0709,
+      "step": 640
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 642,
+  "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": 8.177727396863017e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

114_128_e3_3e-5/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d4241fa3ffe6af265c0232e42de890dbb7ac88a49db2579ee65822e674dba01
+size 8209

114_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)