Upload folder using huggingface_hub

121_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": [
+    "q_proj",
+    "down_proj",
+    "gate_proj",
+    "up_proj",
+    "k_proj",
+    "o_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

121_128_e3_3e-5/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01f0eafb217b5c14600b02a17055aba38e92f29a8c63684d6ea931abc0e21afd
+size 791751704

121_128_e3_3e-5/latest

@@ -0,0 +1 @@
+global_step669

121_128_e3_3e-5/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8e62c8d5d0a922fa0d6d0f3a079ae5bfb3005e6b7f8eece8bc820a75b8a42c5
+size 16389

121_128_e3_3e-5/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1592eeab741d4f8902f393741d65533f5d0e9b6d0bddec268a35b69a033184d6
+size 16389

121_128_e3_3e-5/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e78cdc818339cb1970b9cdb2e17e42bf88f7109fd0ade64ef2c7defb3ace6e27
+size 16389

121_128_e3_3e-5/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb20741c2528c80dd13abe9a5bedab661afca1b1fd656e4082853cbe514c285f
+size 16389

121_128_e3_3e-5/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5082eca4169d0f869888c5f9e990470e01fb1130a8e64244f0ee2339846173c2
+size 16389

121_128_e3_3e-5/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc7a255266ada493277711c44b31bdeba5c31e9b6d326fb1a5256436dc26da05
+size 16389

121_128_e3_3e-5/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0ad1daa3eaaca28a3855bb509d13f2e380899e4e5685b85ab411bd1a3211a357
+size 16389

121_128_e3_3e-5/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae46ae3b66d23ff03cd636ee655373ccd1221d973a8c5191b3af3adbe346e63b
+size 16389

121_128_e3_3e-5/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:526fb229b96c87c191bea76e4a6f87ef3f55b59569a5f5ff72be689b7e409fa7
+size 1401

121_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
+  }
+}

121_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
+}

121_128_e3_3e-5/trainer_state.json

@@ -0,0 +1,965 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 669,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.02242152466367713,
+      "grad_norm": 0.8491255044937134,
+      "learning_rate": 3.5294117647058825e-06,
+      "loss": 1.322,
+      "step": 5
+    },
+    {
+      "epoch": 0.04484304932735426,
+      "grad_norm": 0.663048505783081,
+      "learning_rate": 7.941176470588236e-06,
+      "loss": 1.3167,
+      "step": 10
+    },
+    {
+      "epoch": 0.06726457399103139,
+      "grad_norm": 0.5135542154312134,
+      "learning_rate": 1.2352941176470587e-05,
+      "loss": 1.2641,
+      "step": 15
+    },
+    {
+      "epoch": 0.08968609865470852,
+      "grad_norm": 0.6179513931274414,
+      "learning_rate": 1.6764705882352943e-05,
+      "loss": 1.2789,
+      "step": 20
+    },
+    {
+      "epoch": 0.11210762331838565,
+      "grad_norm": 0.5820193886756897,
+      "learning_rate": 2.1176470588235296e-05,
+      "loss": 1.2185,
+      "step": 25
+    },
+    {
+      "epoch": 0.13452914798206278,
+      "grad_norm": 0.684674859046936,
+      "learning_rate": 2.5588235294117648e-05,
+      "loss": 1.173,
+      "step": 30
+    },
+    {
+      "epoch": 0.15695067264573992,
+      "grad_norm": 0.7267729043960571,
+      "learning_rate": 3e-05,
+      "loss": 1.2027,
+      "step": 35
+    },
+    {
+      "epoch": 0.17937219730941703,
+      "grad_norm": 0.5468352437019348,
+      "learning_rate": 2.9995410858796248e-05,
+      "loss": 1.1577,
+      "step": 40
+    },
+    {
+      "epoch": 0.20179372197309417,
+      "grad_norm": 0.5696616172790527,
+      "learning_rate": 2.9981646243213917e-05,
+      "loss": 1.1024,
+      "step": 45
+    },
+    {
+      "epoch": 0.2242152466367713,
+      "grad_norm": 0.5791175961494446,
+      "learning_rate": 2.995871457562162e-05,
+      "loss": 1.088,
+      "step": 50
+    },
+    {
+      "epoch": 0.24663677130044842,
+      "grad_norm": 0.6192290186882019,
+      "learning_rate": 2.99266298875741e-05,
+      "loss": 0.9442,
+      "step": 55
+    },
+    {
+      "epoch": 0.26905829596412556,
+      "grad_norm": 0.6459205150604248,
+      "learning_rate": 2.9885411811226547e-05,
+      "loss": 1.0248,
+      "step": 60
+    },
+    {
+      "epoch": 0.2914798206278027,
+      "grad_norm": 0.5691049695014954,
+      "learning_rate": 2.983508556732196e-05,
+      "loss": 0.9559,
+      "step": 65
+    },
+    {
+      "epoch": 0.31390134529147984,
+      "grad_norm": 0.7176865339279175,
+      "learning_rate": 2.977568194975895e-05,
+      "loss": 0.9119,
+      "step": 70
+    },
+    {
+      "epoch": 0.336322869955157,
+      "grad_norm": 0.7323933839797974,
+      "learning_rate": 2.970723730674938e-05,
+      "loss": 0.9183,
+      "step": 75
+    },
+    {
+      "epoch": 0.35874439461883406,
+      "grad_norm": 0.7059289216995239,
+      "learning_rate": 2.9629793518577443e-05,
+      "loss": 0.8496,
+      "step": 80
+    },
+    {
+      "epoch": 0.3811659192825112,
+      "grad_norm": 0.7546777725219727,
+      "learning_rate": 2.95433979719737e-05,
+      "loss": 0.8433,
+      "step": 85
+    },
+    {
+      "epoch": 0.40358744394618834,
+      "grad_norm": 0.8598536252975464,
+      "learning_rate": 2.9448103531119858e-05,
+      "loss": 0.8164,
+      "step": 90
+    },
+    {
+      "epoch": 0.4260089686098655,
+      "grad_norm": 0.8127933740615845,
+      "learning_rate": 2.9343968505301914e-05,
+      "loss": 0.8105,
+      "step": 95
+    },
+    {
+      "epoch": 0.4484304932735426,
+      "grad_norm": 0.9225526452064514,
+      "learning_rate": 2.9231056613231567e-05,
+      "loss": 0.7928,
+      "step": 100
+    },
+    {
+      "epoch": 0.47085201793721976,
+      "grad_norm": 1.0847933292388916,
+      "learning_rate": 2.9109436944057662e-05,
+      "loss": 0.7284,
+      "step": 105
+    },
+    {
+      "epoch": 0.49327354260089684,
+      "grad_norm": 1.027095913887024,
+      "learning_rate": 2.8979183915091518e-05,
+      "loss": 0.7118,
+      "step": 110
+    },
+    {
+      "epoch": 0.515695067264574,
+      "grad_norm": 0.992284893989563,
+      "learning_rate": 2.8840377226272102e-05,
+      "loss": 0.7123,
+      "step": 115
+    },
+    {
+      "epoch": 0.5381165919282511,
+      "grad_norm": 1.1309938430786133,
+      "learning_rate": 2.869310181139874e-05,
+      "loss": 0.664,
+      "step": 120
+    },
+    {
+      "epoch": 0.5605381165919282,
+      "grad_norm": 0.9915221929550171,
+      "learning_rate": 2.8537447786161393e-05,
+      "loss": 0.6021,
+      "step": 125
+    },
+    {
+      "epoch": 0.5829596412556054,
+      "grad_norm": 1.2357131242752075,
+      "learning_rate": 2.8373510393000163e-05,
+      "loss": 0.6267,
+      "step": 130
+    },
+    {
+      "epoch": 0.6053811659192825,
+      "grad_norm": 1.2844613790512085,
+      "learning_rate": 2.8201389942827827e-05,
+      "loss": 0.6338,
+      "step": 135
+    },
+    {
+      "epoch": 0.6278026905829597,
+      "grad_norm": 1.211376428604126,
+      "learning_rate": 2.8021191753651025e-05,
+      "loss": 0.6,
+      "step": 140
+    },
+    {
+      "epoch": 0.6502242152466368,
+      "grad_norm": 1.1193534135818481,
+      "learning_rate": 2.783302608612774e-05,
+      "loss": 0.5533,
+      "step": 145
+    },
+    {
+      "epoch": 0.672645739910314,
+      "grad_norm": 1.059785008430481,
+      "learning_rate": 2.7637008076100362e-05,
+      "loss": 0.6067,
+      "step": 150
+    },
+    {
+      "epoch": 0.695067264573991,
+      "grad_norm": 1.0363613367080688,
+      "learning_rate": 2.7433257664145757e-05,
+      "loss": 0.5554,
+      "step": 155
+    },
+    {
+      "epoch": 0.7174887892376681,
+      "grad_norm": 1.0606639385223389,
+      "learning_rate": 2.7221899522185365e-05,
+      "loss": 0.5486,
+      "step": 160
+    },
+    {
+      "epoch": 0.7399103139013453,
+      "grad_norm": 1.2458378076553345,
+      "learning_rate": 2.700306297720026e-05,
+      "loss": 0.5029,
+      "step": 165
+    },
+    {
+      "epoch": 0.7623318385650224,
+      "grad_norm": 1.0505344867706299,
+      "learning_rate": 2.6776881932097826e-05,
+      "loss": 0.5358,
+      "step": 170
+    },
+    {
+      "epoch": 0.7847533632286996,
+      "grad_norm": 1.022645354270935,
+      "learning_rate": 2.654349478377855e-05,
+      "loss": 0.4426,
+      "step": 175
+    },
+    {
+      "epoch": 0.8071748878923767,
+      "grad_norm": 1.164555311203003,
+      "learning_rate": 2.630304433845293e-05,
+      "loss": 0.449,
+      "step": 180
+    },
+    {
+      "epoch": 0.8295964125560538,
+      "grad_norm": 1.1087955236434937,
+      "learning_rate": 2.605567772426046e-05,
+      "loss": 0.4771,
+      "step": 185
+    },
+    {
+      "epoch": 0.852017937219731,
+      "grad_norm": 1.4327459335327148,
+      "learning_rate": 2.5801546301244004e-05,
+      "loss": 0.463,
+      "step": 190
+    },
+    {
+      "epoch": 0.874439461883408,
+      "grad_norm": 1.2425979375839233,
+      "learning_rate": 2.554080556873485e-05,
+      "loss": 0.4041,
+      "step": 195
+    },
+    {
+      "epoch": 0.8968609865470852,
+      "grad_norm": 1.2053102254867554,
+      "learning_rate": 2.5273615070204868e-05,
+      "loss": 0.4331,
+      "step": 200
+    },
+    {
+      "epoch": 0.9192825112107623,
+      "grad_norm": 1.3886176347732544,
+      "learning_rate": 2.500013829564419e-05,
+      "loss": 0.3897,
+      "step": 205
+    },
+    {
+      "epoch": 0.9417040358744395,
+      "grad_norm": 1.3009324073791504,
+      "learning_rate": 2.472054258152408e-05,
+      "loss": 0.436,
+      "step": 210
+    },
+    {
+      "epoch": 0.9641255605381166,
+      "grad_norm": 1.37282395362854,
+      "learning_rate": 2.443499900840614e-05,
+      "loss": 0.3576,
+      "step": 215
+    },
+    {
+      "epoch": 0.9865470852017937,
+      "grad_norm": 1.0965322256088257,
+      "learning_rate": 2.414368229626062e-05,
+      "loss": 0.3658,
+      "step": 220
+    },
+    {
+      "epoch": 1.0089686098654709,
+      "grad_norm": 1.0403130054473877,
+      "learning_rate": 2.384677069755779e-05,
+      "loss": 0.3459,
+      "step": 225
+    },
+    {
+      "epoch": 1.031390134529148,
+      "grad_norm": 1.1162190437316895,
+      "learning_rate": 2.3544445888197855e-05,
+      "loss": 0.2756,
+      "step": 230
+    },
+    {
+      "epoch": 1.053811659192825,
+      "grad_norm": 1.1827653646469116,
+      "learning_rate": 2.323689285634609e-05,
+      "loss": 0.2987,
+      "step": 235
+    },
+    {
+      "epoch": 1.0762331838565022,
+      "grad_norm": 1.3933961391448975,
+      "learning_rate": 2.2924299789241255e-05,
+      "loss": 0.2914,
+      "step": 240
+    },
+    {
+      "epoch": 1.0986547085201794,
+      "grad_norm": 1.308779001235962,
+      "learning_rate": 2.2606857958046597e-05,
+      "loss": 0.2821,
+      "step": 245
+    },
+    {
+      "epoch": 1.1210762331838564,
+      "grad_norm": 1.1784348487854004,
+      "learning_rate": 2.2284761600813755e-05,
+      "loss": 0.2597,
+      "step": 250
+    },
+    {
+      "epoch": 1.1434977578475336,
+      "grad_norm": 1.2625153064727783,
+      "learning_rate": 2.1958207803631343e-05,
+      "loss": 0.271,
+      "step": 255
+    },
+    {
+      "epoch": 1.1659192825112108,
+      "grad_norm": 1.319551944732666,
+      "learning_rate": 2.1627396380030803e-05,
+      "loss": 0.2557,
+      "step": 260
+    },
+    {
+      "epoch": 1.188340807174888,
+      "grad_norm": 1.422255039215088,
+      "learning_rate": 2.1292529748723436e-05,
+      "loss": 0.2618,
+      "step": 265
+    },
+    {
+      "epoch": 1.210762331838565,
+      "grad_norm": 1.1897157430648804,
+      "learning_rate": 2.095381280974331e-05,
+      "loss": 0.2707,
+      "step": 270
+    },
+    {
+      "epoch": 1.2331838565022422,
+      "grad_norm": 1.0201008319854736,
+      "learning_rate": 2.06114528190719e-05,
+      "loss": 0.2682,
+      "step": 275
+    },
+    {
+      "epoch": 1.2556053811659194,
+      "grad_norm": 1.6869356632232666,
+      "learning_rate": 2.0265659261821163e-05,
+      "loss": 0.2567,
+      "step": 280
+    },
+    {
+      "epoch": 1.2780269058295963,
+      "grad_norm": 1.0875405073165894,
+      "learning_rate": 1.991664372405265e-05,
+      "loss": 0.2447,
+      "step": 285
+    },
+    {
+      "epoch": 1.3004484304932735,
+      "grad_norm": 1.1517951488494873,
+      "learning_rate": 1.9564619763311043e-05,
+      "loss": 0.2448,
+      "step": 290
+    },
+    {
+      "epoch": 1.3228699551569507,
+      "grad_norm": 1.652977466583252,
+      "learning_rate": 1.92098027779514e-05,
+      "loss": 0.2303,
+      "step": 295
+    },
+    {
+      "epoch": 1.3452914798206277,
+      "grad_norm": 1.0147415399551392,
+      "learning_rate": 1.885240987534002e-05,
+      "loss": 0.2313,
+      "step": 300
+    },
+    {
+      "epoch": 1.3677130044843049,
+      "grad_norm": 1.2878891229629517,
+      "learning_rate": 1.849265973900962e-05,
+      "loss": 0.2364,
+      "step": 305
+    },
+    {
+      "epoch": 1.390134529147982,
+      "grad_norm": 1.0258630514144897,
+      "learning_rate": 1.813077249485003e-05,
+      "loss": 0.1764,
+      "step": 310
+    },
+    {
+      "epoch": 1.4125560538116593,
+      "grad_norm": 1.3200815916061401,
+      "learning_rate": 1.776696957641634e-05,
+      "loss": 0.2022,
+      "step": 315
+    },
+    {
+      "epoch": 1.4349775784753362,
+      "grad_norm": 1.2388659715652466,
+      "learning_rate": 1.7401473589436966e-05,
+      "loss": 0.2014,
+      "step": 320
+    },
+    {
+      "epoch": 1.4573991031390134,
+      "grad_norm": 1.326780080795288,
+      "learning_rate": 1.7034508175604383e-05,
+      "loss": 0.2114,
+      "step": 325
+    },
+    {
+      "epoch": 1.4798206278026906,
+      "grad_norm": 1.1960903406143188,
+      "learning_rate": 1.6666297875732072e-05,
+      "loss": 0.2126,
+      "step": 330
+    },
+    {
+      "epoch": 1.5022421524663678,
+      "grad_norm": 1.134148359298706,
+      "learning_rate": 1.6297067992361185e-05,
+      "loss": 0.1957,
+      "step": 335
+    },
+    {
+      "epoch": 1.5246636771300448,
+      "grad_norm": 1.267330288887024,
+      "learning_rate": 1.5927044451901265e-05,
+      "loss": 0.2024,
+      "step": 340
+    },
+    {
+      "epoch": 1.547085201793722,
+      "grad_norm": 1.1064133644104004,
+      "learning_rate": 1.5556453666389077e-05,
+      "loss": 0.1639,
+      "step": 345
+    },
+    {
+      "epoch": 1.5695067264573992,
+      "grad_norm": 1.1145753860473633,
+      "learning_rate": 1.5185522394950444e-05,
+      "loss": 0.1821,
+      "step": 350
+    },
+    {
+      "epoch": 1.5919282511210762,
+      "grad_norm": 1.2523396015167236,
+      "learning_rate": 1.4814477605049557e-05,
+      "loss": 0.161,
+      "step": 355
+    },
+    {
+      "epoch": 1.6143497757847534,
+      "grad_norm": 1.1929575204849243,
+      "learning_rate": 1.4443546333610923e-05,
+      "loss": 0.1689,
+      "step": 360
+    },
+    {
+      "epoch": 1.6367713004484306,
+      "grad_norm": 1.2169147729873657,
+      "learning_rate": 1.4072955548098741e-05,
+      "loss": 0.1667,
+      "step": 365
+    },
+    {
+      "epoch": 1.6591928251121075,
+      "grad_norm": 0.9864022731781006,
+      "learning_rate": 1.3702932007638819e-05,
+      "loss": 0.1536,
+      "step": 370
+    },
+    {
+      "epoch": 1.6816143497757847,
+      "grad_norm": 1.1148977279663086,
+      "learning_rate": 1.3333702124267932e-05,
+      "loss": 0.1457,
+      "step": 375
+    },
+    {
+      "epoch": 1.704035874439462,
+      "grad_norm": 1.037030577659607,
+      "learning_rate": 1.2965491824395614e-05,
+      "loss": 0.1721,
+      "step": 380
+    },
+    {
+      "epoch": 1.726457399103139,
+      "grad_norm": 1.2554570436477661,
+      "learning_rate": 1.2598526410563037e-05,
+      "loss": 0.1567,
+      "step": 385
+    },
+    {
+      "epoch": 1.7488789237668163,
+      "grad_norm": 1.038957118988037,
+      "learning_rate": 1.2233030423583662e-05,
+      "loss": 0.1349,
+      "step": 390
+    },
+    {
+      "epoch": 1.7713004484304933,
+      "grad_norm": 1.0190961360931396,
+      "learning_rate": 1.1869227505149974e-05,
+      "loss": 0.1301,
+      "step": 395
+    },
+    {
+      "epoch": 1.7937219730941703,
+      "grad_norm": 1.1681493520736694,
+      "learning_rate": 1.150734026099038e-05,
+      "loss": 0.1558,
+      "step": 400
+    },
+    {
+      "epoch": 1.8161434977578477,
+      "grad_norm": 1.213783621788025,
+      "learning_rate": 1.1147590124659983e-05,
+      "loss": 0.1214,
+      "step": 405
+    },
+    {
+      "epoch": 1.8385650224215246,
+      "grad_norm": 0.9164388179779053,
+      "learning_rate": 1.0790197222048609e-05,
+      "loss": 0.1336,
+      "step": 410
+    },
+    {
+      "epoch": 1.8609865470852018,
+      "grad_norm": 1.075270652770996,
+      "learning_rate": 1.0435380236688958e-05,
+      "loss": 0.1377,
+      "step": 415
+    },
+    {
+      "epoch": 1.883408071748879,
+      "grad_norm": 1.1690846681594849,
+      "learning_rate": 1.0083356275947348e-05,
+      "loss": 0.132,
+      "step": 420
+    },
+    {
+      "epoch": 1.905829596412556,
+      "grad_norm": 1.0824881792068481,
+      "learning_rate": 9.734340738178836e-06,
+      "loss": 0.124,
+      "step": 425
+    },
+    {
+      "epoch": 1.9282511210762332,
+      "grad_norm": 1.0719696283340454,
+      "learning_rate": 9.388547180928103e-06,
+      "loss": 0.112,
+      "step": 430
+    },
+    {
+      "epoch": 1.9506726457399104,
+      "grad_norm": 0.9826474785804749,
+      "learning_rate": 9.046187190256694e-06,
+      "loss": 0.1162,
+      "step": 435
+    },
+    {
+      "epoch": 1.9730941704035874,
+      "grad_norm": 0.9230247139930725,
+      "learning_rate": 8.707470251276568e-06,
+      "loss": 0.107,
+      "step": 440
+    },
+    {
+      "epoch": 1.9955156950672646,
+      "grad_norm": 0.9789291024208069,
+      "learning_rate": 8.372603619969201e-06,
+      "loss": 0.1181,
+      "step": 445
+    },
+    {
+      "epoch": 2.0179372197309418,
+      "grad_norm": 0.9269797801971436,
+      "learning_rate": 8.04179219636866e-06,
+      "loss": 0.0855,
+      "step": 450
+    },
+    {
+      "epoch": 2.0403587443946187,
+      "grad_norm": 0.7365989685058594,
+      "learning_rate": 7.715238399186246e-06,
+      "loss": 0.0921,
+      "step": 455
+    },
+    {
+      "epoch": 2.062780269058296,
+      "grad_norm": 1.0451802015304565,
+      "learning_rate": 7.393142041953403e-06,
+      "loss": 0.0962,
+      "step": 460
+    },
+    {
+      "epoch": 2.085201793721973,
+      "grad_norm": 0.8351704478263855,
+      "learning_rate": 7.075700210758744e-06,
+      "loss": 0.0989,
+      "step": 465
+    },
+    {
+      "epoch": 2.10762331838565,
+      "grad_norm": 0.8635206818580627,
+      "learning_rate": 6.763107143653917e-06,
+      "loss": 0.078,
+      "step": 470
+    },
+    {
+      "epoch": 2.1300448430493275,
+      "grad_norm": 0.8441980481147766,
+      "learning_rate": 6.455554111802146e-06,
+      "loss": 0.0891,
+      "step": 475
+    },
+    {
+      "epoch": 2.1524663677130045,
+      "grad_norm": 0.876447856426239,
+      "learning_rate": 6.153229302442209e-06,
+      "loss": 0.0788,
+      "step": 480
+    },
+    {
+      "epoch": 2.1748878923766815,
+      "grad_norm": 0.9017621874809265,
+      "learning_rate": 5.856317703739383e-06,
+      "loss": 0.0852,
+      "step": 485
+    },
+    {
+      "epoch": 2.197309417040359,
+      "grad_norm": 0.7752906680107117,
+      "learning_rate": 5.5650009915938675e-06,
+      "loss": 0.0761,
+      "step": 490
+    },
+    {
+      "epoch": 2.219730941704036,
+      "grad_norm": 0.7093114852905273,
+      "learning_rate": 5.279457418475921e-06,
+      "loss": 0.0721,
+      "step": 495
+    },
+    {
+      "epoch": 2.242152466367713,
+      "grad_norm": 0.9206439256668091,
+      "learning_rate": 4.999861704355809e-06,
+      "loss": 0.0842,
+      "step": 500
+    },
+    {
+      "epoch": 2.2645739910313902,
+      "grad_norm": 1.008296012878418,
+      "learning_rate": 4.726384929795133e-06,
+      "loss": 0.0848,
+      "step": 505
+    },
+    {
+      "epoch": 2.286995515695067,
+      "grad_norm": 0.7057260870933533,
+      "learning_rate": 4.45919443126515e-06,
+      "loss": 0.0862,
+      "step": 510
+    },
+    {
+      "epoch": 2.3094170403587446,
+      "grad_norm": 0.7205430269241333,
+      "learning_rate": 4.198453698755995e-06,
+      "loss": 0.0714,
+      "step": 515
+    },
+    {
+      "epoch": 2.3318385650224216,
+      "grad_norm": 0.7457002401351929,
+      "learning_rate": 3.944322275739544e-06,
+      "loss": 0.0705,
+      "step": 520
+    },
+    {
+      "epoch": 2.3542600896860986,
+      "grad_norm": 0.7221155762672424,
+      "learning_rate": 3.6969556615470684e-06,
+      "loss": 0.0803,
+      "step": 525
+    },
+    {
+      "epoch": 2.376681614349776,
+      "grad_norm": 0.8441908359527588,
+      "learning_rate": 3.4565052162214573e-06,
+      "loss": 0.0802,
+      "step": 530
+    },
+    {
+      "epoch": 2.399103139013453,
+      "grad_norm": 0.6702395081520081,
+      "learning_rate": 3.2231180679021766e-06,
+      "loss": 0.0777,
+      "step": 535
+    },
+    {
+      "epoch": 2.42152466367713,
+      "grad_norm": 0.6875882148742676,
+      "learning_rate": 2.9969370227997434e-06,
+      "loss": 0.0756,
+      "step": 540
+    },
+    {
+      "epoch": 2.4439461883408073,
+      "grad_norm": 0.6272410750389099,
+      "learning_rate": 2.7781004778146357e-06,
+      "loss": 0.0741,
+      "step": 545
+    },
+    {
+      "epoch": 2.4663677130044843,
+      "grad_norm": 0.8176401853561401,
+      "learning_rate": 2.566742335854244e-06,
+      "loss": 0.0746,
+      "step": 550
+    },
+    {
+      "epoch": 2.4887892376681613,
+      "grad_norm": 0.631503701210022,
+      "learning_rate": 2.3629919238996393e-06,
+      "loss": 0.0738,
+      "step": 555
+    },
+    {
+      "epoch": 2.5112107623318387,
+      "grad_norm": 0.636724054813385,
+      "learning_rate": 2.1669739138722617e-06,
+      "loss": 0.0741,
+      "step": 560
+    },
+    {
+      "epoch": 2.5336322869955157,
+      "grad_norm": 0.7829221487045288,
+      "learning_rate": 1.9788082463489743e-06,
+      "loss": 0.0725,
+      "step": 565
+    },
+    {
+      "epoch": 2.5560538116591927,
+      "grad_norm": 0.7489262223243713,
+      "learning_rate": 1.7986100571721774e-06,
+      "loss": 0.0642,
+      "step": 570
+    },
+    {
+      "epoch": 2.57847533632287,
+      "grad_norm": 0.6036618947982788,
+      "learning_rate": 1.6264896069998337e-06,
+      "loss": 0.0649,
+      "step": 575
+    },
+    {
+      "epoch": 2.600896860986547,
+      "grad_norm": 0.7053378224372864,
+      "learning_rate": 1.462552213838605e-06,
+      "loss": 0.0623,
+      "step": 580
+    },
+    {
+      "epoch": 2.623318385650224,
+      "grad_norm": 0.6539296507835388,
+      "learning_rate": 1.3068981886012627e-06,
+      "loss": 0.0609,
+      "step": 585
+    },
+    {
+      "epoch": 2.6457399103139014,
+      "grad_norm": 0.617291271686554,
+      "learning_rate": 1.1596227737279014e-06,
+      "loss": 0.0627,
+      "step": 590
+    },
+    {
+      "epoch": 2.6681614349775784,
+      "grad_norm": 0.706615686416626,
+      "learning_rate": 1.020816084908483e-06,
+      "loss": 0.0766,
+      "step": 595
+    },
+    {
+      "epoch": 2.6905829596412554,
+      "grad_norm": 0.6039079427719116,
+      "learning_rate": 8.905630559423428e-07,
+      "loss": 0.077,
+      "step": 600
+    },
+    {
+      "epoch": 2.713004484304933,
+      "grad_norm": 0.6305246353149414,
+      "learning_rate": 7.689433867684359e-07,
+      "loss": 0.0724,
+      "step": 605
+    },
+    {
+      "epoch": 2.7354260089686098,
+      "grad_norm": 0.7037762999534607,
+      "learning_rate": 6.560314946980889e-07,
+      "loss": 0.0619,
+      "step": 610
+    },
+    {
+      "epoch": 2.7578475336322867,
+      "grad_norm": 0.6866577863693237,
+      "learning_rate": 5.518964688801436e-07,
+      "loss": 0.0579,
+      "step": 615
+    },
+    {
+      "epoch": 2.780269058295964,
+      "grad_norm": 0.6411738395690918,
+      "learning_rate": 4.566020280263006e-07,
+      "loss": 0.0682,
+      "step": 620
+    },
+    {
+      "epoch": 2.802690582959641,
+      "grad_norm": 0.6658198237419128,
+      "learning_rate": 3.702064814225609e-07,
+      "loss": 0.0668,
+      "step": 625
+    },
+    {
+      "epoch": 2.8251121076233185,
+      "grad_norm": 0.6498026251792908,
+      "learning_rate": 2.9276269325062e-07,
+      "loss": 0.0694,
+      "step": 630
+    },
+    {
+      "epoch": 2.8475336322869955,
+      "grad_norm": 0.6728014945983887,
+      "learning_rate": 2.243180502410508e-07,
+      "loss": 0.0628,
+      "step": 635
+    },
+    {
+      "epoch": 2.8699551569506725,
+      "grad_norm": 0.6942362785339355,
+      "learning_rate": 1.6491443267803886e-07,
+      "loss": 0.0634,
+      "step": 640
+    },
+    {
+      "epoch": 2.89237668161435,
+      "grad_norm": 0.6509578227996826,
+      "learning_rate": 1.1458818877345622e-07,
+      "loss": 0.0755,
+      "step": 645
+    },
+    {
+      "epoch": 2.914798206278027,
+      "grad_norm": 0.6404685974121094,
+      "learning_rate": 7.33701124259023e-08,
+      "loss": 0.0632,
+      "step": 650
+    },
+    {
+      "epoch": 2.9372197309417043,
+      "grad_norm": 0.7250242233276367,
+      "learning_rate": 4.1285424378379456e-08,
+      "loss": 0.0614,
+      "step": 655
+    },
+    {
+      "epoch": 2.9596412556053813,
+      "grad_norm": 0.6524393558502197,
+      "learning_rate": 1.8353756786080865e-08,
+      "loss": 0.0619,
+      "step": 660
+    },
+    {
+      "epoch": 2.9820627802690582,
+      "grad_norm": 0.58048015832901,
+      "learning_rate": 4.589141203752978e-09,
+      "loss": 0.0647,
+      "step": 665
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 669,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 2000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 9.337086305558856e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

121_128_e3_3e-5/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8139c964ff034a69571dddea2d8032e90aac1864d8026bbb510703879061e20
+size 8209

121_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)