Upload folder using huggingface_hub

125_128_e3_3e-5/adapter_config.json

@@ -0,0 +1,39 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "ibm-granite/granite-3.3-8b-base",
+  "bias": "none",
+  "corda_config": null,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 256,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 128,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "up_proj",
+    "k_proj",
+    "q_proj",
+    "down_proj",
+    "o_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

125_128_e3_3e-5/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b109f76c0e6993f2d24b25710871b37767299d734efee53834e96730342cd31
+size 791751704

125_128_e3_3e-5/latest

@@ -0,0 +1 @@
+global_step636

125_128_e3_3e-5/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1dfb24079690a76eae8020b5df95f319ba5921e96460f26ae8334674d17310c6
+size 16389

125_128_e3_3e-5/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:847884959f4bd4af051fc9c6c9fbc837d4d01da5ca92bccd78b45ef1aa8c3da7
+size 16389

125_128_e3_3e-5/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f39872f016a9298f6d0aec747dae03038840fff8c1acdb718bf87ac3468bdf19
+size 16389

125_128_e3_3e-5/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f45e0114dd8872e15ad23c8bd36b92336c875056b10cbb6f81f641f7e7d1a446
+size 16389

125_128_e3_3e-5/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e10f790fa744eb1e4541f3ebed4886bb416fc423d02e7b2881b100d19256bcb
+size 16389

125_128_e3_3e-5/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac2bd2f682ed6626040137bd4450eccd0aa6d0e74c9ae0f92956f67a229790f5
+size 16389

125_128_e3_3e-5/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6bde4dc355c9fe0deeb7b0ed3ca547424569a084b6af78914e80ec379a5efce0
+size 16389

125_128_e3_3e-5/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed4e801f3c8e5e9fbc4c790a22e3d065bc3179e003404a3e08bccbd50179e584
+size 16389

125_128_e3_3e-5/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:212bb4b2d133983fa20226ade73c89a80bf24adcb0ddd976df52f966359503bc
+size 1401

125_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
+  }
+}

125_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
+}

125_128_e3_3e-5/trainer_state.json

@@ -0,0 +1,923 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 636,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.02358490566037736,
+      "grad_norm": 1.0230686664581299,
+      "learning_rate": 3.75e-06,
+      "loss": 1.4908,
+      "step": 5
+    },
+    {
+      "epoch": 0.04716981132075472,
+      "grad_norm": 0.6779772639274597,
+      "learning_rate": 8.4375e-06,
+      "loss": 1.44,
+      "step": 10
+    },
+    {
+      "epoch": 0.07075471698113207,
+      "grad_norm": 0.5684516429901123,
+      "learning_rate": 1.3125e-05,
+      "loss": 1.4457,
+      "step": 15
+    },
+    {
+      "epoch": 0.09433962264150944,
+      "grad_norm": 0.6403588652610779,
+      "learning_rate": 1.78125e-05,
+      "loss": 1.4014,
+      "step": 20
+    },
+    {
+      "epoch": 0.1179245283018868,
+      "grad_norm": 0.5737028121948242,
+      "learning_rate": 2.25e-05,
+      "loss": 1.3248,
+      "step": 25
+    },
+    {
+      "epoch": 0.14150943396226415,
+      "grad_norm": 0.525193989276886,
+      "learning_rate": 2.71875e-05,
+      "loss": 1.3713,
+      "step": 30
+    },
+    {
+      "epoch": 0.1650943396226415,
+      "grad_norm": 0.6601390242576599,
+      "learning_rate": 2.9999188397817107e-05,
+      "loss": 1.2913,
+      "step": 35
+    },
+    {
+      "epoch": 0.18867924528301888,
+      "grad_norm": 0.6258649826049805,
+      "learning_rate": 2.9990058881861022e-05,
+      "loss": 1.2302,
+      "step": 40
+    },
+    {
+      "epoch": 0.21226415094339623,
+      "grad_norm": 0.6430497765541077,
+      "learning_rate": 2.9970791542124925e-05,
+      "loss": 1.1851,
+      "step": 45
+    },
+    {
+      "epoch": 0.2358490566037736,
+      "grad_norm": 0.6849141120910645,
+      "learning_rate": 2.9941399409171037e-05,
+      "loss": 1.2004,
+      "step": 50
+    },
+    {
+      "epoch": 0.25943396226415094,
+      "grad_norm": 0.6358975768089294,
+      "learning_rate": 2.990190236099094e-05,
+      "loss": 1.0782,
+      "step": 55
+    },
+    {
+      "epoch": 0.2830188679245283,
+      "grad_norm": 0.7141816020011902,
+      "learning_rate": 2.985232710956206e-05,
+      "loss": 1.1762,
+      "step": 60
+    },
+    {
+      "epoch": 0.30660377358490565,
+      "grad_norm": 0.6631730794906616,
+      "learning_rate": 2.979270718278224e-05,
+      "loss": 1.0616,
+      "step": 65
+    },
+    {
+      "epoch": 0.330188679245283,
+      "grad_norm": 0.7165650725364685,
+      "learning_rate": 2.972308290179471e-05,
+      "loss": 1.0715,
+      "step": 70
+    },
+    {
+      "epoch": 0.35377358490566035,
+      "grad_norm": 0.7615548372268677,
+      "learning_rate": 2.964350135371883e-05,
+      "loss": 1.0452,
+      "step": 75
+    },
+    {
+      "epoch": 0.37735849056603776,
+      "grad_norm": 0.878010094165802,
+      "learning_rate": 2.9554016359804843e-05,
+      "loss": 0.9841,
+      "step": 80
+    },
+    {
+      "epoch": 0.4009433962264151,
+      "grad_norm": 0.8644310832023621,
+      "learning_rate": 2.9454688439034475e-05,
+      "loss": 0.9674,
+      "step": 85
+    },
+    {
+      "epoch": 0.42452830188679247,
+      "grad_norm": 0.9034149646759033,
+      "learning_rate": 2.9345584767191685e-05,
+      "loss": 0.9373,
+      "step": 90
+    },
+    {
+      "epoch": 0.4481132075471698,
+      "grad_norm": 0.9173258543014526,
+      "learning_rate": 2.9226779131431533e-05,
+      "loss": 0.8683,
+      "step": 95
+    },
+    {
+      "epoch": 0.4716981132075472,
+      "grad_norm": 1.0164599418640137,
+      "learning_rate": 2.909835188037764e-05,
+      "loss": 0.8595,
+      "step": 100
+    },
+    {
+      "epoch": 0.49528301886792453,
+      "grad_norm": 1.2891350984573364,
+      "learning_rate": 2.896038986978223e-05,
+      "loss": 0.8334,
+      "step": 105
+    },
+    {
+      "epoch": 0.5188679245283019,
+      "grad_norm": 0.9736058115959167,
+      "learning_rate": 2.8812986403785236e-05,
+      "loss": 0.8018,
+      "step": 110
+    },
+    {
+      "epoch": 0.5424528301886793,
+      "grad_norm": 1.0428775548934937,
+      "learning_rate": 2.8656241171812475e-05,
+      "loss": 0.798,
+      "step": 115
+    },
+    {
+      "epoch": 0.5660377358490566,
+      "grad_norm": 1.1972447633743286,
+      "learning_rate": 2.8490260181155358e-05,
+      "loss": 0.715,
+      "step": 120
+    },
+    {
+      "epoch": 0.589622641509434,
+      "grad_norm": 1.1746220588684082,
+      "learning_rate": 2.8315155685277816e-05,
+      "loss": 0.7373,
+      "step": 125
+    },
+    {
+      "epoch": 0.6132075471698113,
+      "grad_norm": 1.3712944984436035,
+      "learning_rate": 2.813104610789895e-05,
+      "loss": 0.6915,
+      "step": 130
+    },
+    {
+      "epoch": 0.6367924528301887,
+      "grad_norm": 1.3445384502410889,
+      "learning_rate": 2.7938055962902694e-05,
+      "loss": 0.7224,
+      "step": 135
+    },
+    {
+      "epoch": 0.660377358490566,
+      "grad_norm": 1.246478796005249,
+      "learning_rate": 2.7736315770128663e-05,
+      "loss": 0.6648,
+      "step": 140
+    },
+    {
+      "epoch": 0.6839622641509434,
+      "grad_norm": 1.2918355464935303,
+      "learning_rate": 2.7525961967101216e-05,
+      "loss": 0.6263,
+      "step": 145
+    },
+    {
+      "epoch": 0.7075471698113207,
+      "grad_norm": 1.2422397136688232,
+      "learning_rate": 2.7307136816756293e-05,
+      "loss": 0.6348,
+      "step": 150
+    },
+    {
+      "epoch": 0.7311320754716981,
+      "grad_norm": 1.0136449337005615,
+      "learning_rate": 2.7079988311228567e-05,
+      "loss": 0.6012,
+      "step": 155
+    },
+    {
+      "epoch": 0.7547169811320755,
+      "grad_norm": 1.30122709274292,
+      "learning_rate": 2.6844670071763906e-05,
+      "loss": 0.5677,
+      "step": 160
+    },
+    {
+      "epoch": 0.7783018867924528,
+      "grad_norm": 1.219581127166748,
+      "learning_rate": 2.660134124482482e-05,
+      "loss": 0.5234,
+      "step": 165
+    },
+    {
+      "epoch": 0.8018867924528302,
+      "grad_norm": 1.2171543836593628,
+      "learning_rate": 2.635016639445922e-05,
+      "loss": 0.5706,
+      "step": 170
+    },
+    {
+      "epoch": 0.8254716981132075,
+      "grad_norm": 1.3339793682098389,
+      "learning_rate": 2.6091315391005255e-05,
+      "loss": 0.4864,
+      "step": 175
+    },
+    {
+      "epoch": 0.8490566037735849,
+      "grad_norm": 1.1416600942611694,
+      "learning_rate": 2.5824963296207435e-05,
+      "loss": 0.4808,
+      "step": 180
+    },
+    {
+      "epoch": 0.8726415094339622,
+      "grad_norm": 1.3225994110107422,
+      "learning_rate": 2.5551290244821856e-05,
+      "loss": 0.5098,
+      "step": 185
+    },
+    {
+      "epoch": 0.8962264150943396,
+      "grad_norm": 1.7722339630126953,
+      "learning_rate": 2.527048132279048e-05,
+      "loss": 0.4977,
+      "step": 190
+    },
+    {
+      "epoch": 0.9198113207547169,
+      "grad_norm": 1.4475733041763306,
+      "learning_rate": 2.4982726442066954e-05,
+      "loss": 0.4651,
+      "step": 195
+    },
+    {
+      "epoch": 0.9433962264150944,
+      "grad_norm": 1.4808722734451294,
+      "learning_rate": 2.4688220212178528e-05,
+      "loss": 0.4677,
+      "step": 200
+    },
+    {
+      "epoch": 0.9669811320754716,
+      "grad_norm": 1.4779961109161377,
+      "learning_rate": 2.438716180861106e-05,
+      "loss": 0.4563,
+      "step": 205
+    },
+    {
+      "epoch": 0.9905660377358491,
+      "grad_norm": 1.3232123851776123,
+      "learning_rate": 2.4079754838105993e-05,
+      "loss": 0.4168,
+      "step": 210
+    },
+    {
+      "epoch": 1.0141509433962264,
+      "grad_norm": 1.6339937448501587,
+      "learning_rate": 2.3766207200960456e-05,
+      "loss": 0.4047,
+      "step": 215
+    },
+    {
+      "epoch": 1.0377358490566038,
+      "grad_norm": 1.3375585079193115,
+      "learning_rate": 2.3446730950423685e-05,
+      "loss": 0.3694,
+      "step": 220
+    },
+    {
+      "epoch": 1.0613207547169812,
+      "grad_norm": 1.6592459678649902,
+      "learning_rate": 2.3121542149284712e-05,
+      "loss": 0.3247,
+      "step": 225
+    },
+    {
+      "epoch": 1.0849056603773586,
+      "grad_norm": 1.3480408191680908,
+      "learning_rate": 2.2790860723748443e-05,
+      "loss": 0.2963,
+      "step": 230
+    },
+    {
+      "epoch": 1.1084905660377358,
+      "grad_norm": 1.2334380149841309,
+      "learning_rate": 2.245491031469887e-05,
+      "loss": 0.3112,
+      "step": 235
+    },
+    {
+      "epoch": 1.1320754716981132,
+      "grad_norm": 1.4530185461044312,
+      "learning_rate": 2.2113918126450036e-05,
+      "loss": 0.3086,
+      "step": 240
+    },
+    {
+      "epoch": 1.1556603773584906,
+      "grad_norm": 1.3657845258712769,
+      "learning_rate": 2.1768114773087063e-05,
+      "loss": 0.2966,
+      "step": 245
+    },
+    {
+      "epoch": 1.179245283018868,
+      "grad_norm": 1.5218877792358398,
+      "learning_rate": 2.141773412250109e-05,
+      "loss": 0.3193,
+      "step": 250
+    },
+    {
+      "epoch": 1.2028301886792452,
+      "grad_norm": 1.424222707748413,
+      "learning_rate": 2.106301313822373e-05,
+      "loss": 0.3167,
+      "step": 255
+    },
+    {
+      "epoch": 1.2264150943396226,
+      "grad_norm": 1.4472578763961792,
+      "learning_rate": 2.0704191719167848e-05,
+      "loss": 0.3534,
+      "step": 260
+    },
+    {
+      "epoch": 1.25,
+      "grad_norm": 1.3955225944519043,
+      "learning_rate": 2.0341512537383202e-05,
+      "loss": 0.2728,
+      "step": 265
+    },
+    {
+      "epoch": 1.2735849056603774,
+      "grad_norm": 1.3127011060714722,
+      "learning_rate": 1.9975220873936587e-05,
+      "loss": 0.2685,
+      "step": 270
+    },
+    {
+      "epoch": 1.2971698113207548,
+      "grad_norm": 1.2843509912490845,
+      "learning_rate": 1.960556445302743e-05,
+      "loss": 0.2864,
+      "step": 275
+    },
+    {
+      "epoch": 1.320754716981132,
+      "grad_norm": 1.2289397716522217,
+      "learning_rate": 1.9232793274451188e-05,
+      "loss": 0.2494,
+      "step": 280
+    },
+    {
+      "epoch": 1.3443396226415094,
+      "grad_norm": 1.1893107891082764,
+      "learning_rate": 1.8857159444523688e-05,
+      "loss": 0.2451,
+      "step": 285
+    },
+    {
+      "epoch": 1.3679245283018868,
+      "grad_norm": 1.600691318511963,
+      "learning_rate": 1.847891700558083e-05,
+      "loss": 0.2874,
+      "step": 290
+    },
+    {
+      "epoch": 1.3915094339622642,
+      "grad_norm": 1.2918438911437988,
+      "learning_rate": 1.8098321764168964e-05,
+      "loss": 0.273,
+      "step": 295
+    },
+    {
+      "epoch": 1.4150943396226414,
+      "grad_norm": 1.4190521240234375,
+      "learning_rate": 1.771563111804211e-05,
+      "loss": 0.2568,
+      "step": 300
+    },
+    {
+      "epoch": 1.4386792452830188,
+      "grad_norm": 1.3648803234100342,
+      "learning_rate": 1.7331103882083075e-05,
+      "loss": 0.2227,
+      "step": 305
+    },
+    {
+      "epoch": 1.4622641509433962,
+      "grad_norm": 1.39310884475708,
+      "learning_rate": 1.6945000113266114e-05,
+      "loss": 0.2551,
+      "step": 310
+    },
+    {
+      "epoch": 1.4858490566037736,
+      "grad_norm": 1.7713751792907715,
+      "learning_rate": 1.6557580934779605e-05,
+      "loss": 0.2512,
+      "step": 315
+    },
+    {
+      "epoch": 1.509433962264151,
+      "grad_norm": 1.499708652496338,
+      "learning_rate": 1.6169108359427585e-05,
+      "loss": 0.2312,
+      "step": 320
+    },
+    {
+      "epoch": 1.5330188679245285,
+      "grad_norm": 1.5003621578216553,
+      "learning_rate": 1.5779845112429706e-05,
+      "loss": 0.2086,
+      "step": 325
+    },
+    {
+      "epoch": 1.5566037735849056,
+      "grad_norm": 1.1198595762252808,
+      "learning_rate": 1.5390054453739335e-05,
+      "loss": 0.1988,
+      "step": 330
+    },
+    {
+      "epoch": 1.580188679245283,
+      "grad_norm": 1.3855667114257812,
+      "learning_rate": 1.5e-05,
+      "loss": 0.1953,
+      "step": 335
+    },
+    {
+      "epoch": 1.6037735849056602,
+      "grad_norm": 1.267496109008789,
+      "learning_rate": 1.4609945546260669e-05,
+      "loss": 0.2054,
+      "step": 340
+    },
+    {
+      "epoch": 1.6273584905660377,
+      "grad_norm": 1.2617906332015991,
+      "learning_rate": 1.4220154887570298e-05,
+      "loss": 0.1771,
+      "step": 345
+    },
+    {
+      "epoch": 1.650943396226415,
+      "grad_norm": 1.174225091934204,
+      "learning_rate": 1.383089164057242e-05,
+      "loss": 0.1799,
+      "step": 350
+    },
+    {
+      "epoch": 1.6745283018867925,
+      "grad_norm": 1.3432834148406982,
+      "learning_rate": 1.34424190652204e-05,
+      "loss": 0.1983,
+      "step": 355
+    },
+    {
+      "epoch": 1.6981132075471699,
+      "grad_norm": 1.215848684310913,
+      "learning_rate": 1.3054999886733883e-05,
+      "loss": 0.1772,
+      "step": 360
+    },
+    {
+      "epoch": 1.7216981132075473,
+      "grad_norm": 1.1607779264450073,
+      "learning_rate": 1.2668896117916928e-05,
+      "loss": 0.1888,
+      "step": 365
+    },
+    {
+      "epoch": 1.7452830188679245,
+      "grad_norm": 1.5456607341766357,
+      "learning_rate": 1.2284368881957892e-05,
+      "loss": 0.1675,
+      "step": 370
+    },
+    {
+      "epoch": 1.7688679245283019,
+      "grad_norm": 1.4072362184524536,
+      "learning_rate": 1.190167823583104e-05,
+      "loss": 0.1734,
+      "step": 375
+    },
+    {
+      "epoch": 1.7924528301886793,
+      "grad_norm": 1.2130942344665527,
+      "learning_rate": 1.1521082994419172e-05,
+      "loss": 0.1608,
+      "step": 380
+    },
+    {
+      "epoch": 1.8160377358490565,
+      "grad_norm": 1.6064542531967163,
+      "learning_rate": 1.1142840555476313e-05,
+      "loss": 0.1676,
+      "step": 385
+    },
+    {
+      "epoch": 1.8396226415094339,
+      "grad_norm": 1.285561442375183,
+      "learning_rate": 1.0767206725548813e-05,
+      "loss": 0.1544,
+      "step": 390
+    },
+    {
+      "epoch": 1.8632075471698113,
+      "grad_norm": 1.398463249206543,
+      "learning_rate": 1.0394435546972577e-05,
+      "loss": 0.1512,
+      "step": 395
+    },
+    {
+      "epoch": 1.8867924528301887,
+      "grad_norm": 1.0858711004257202,
+      "learning_rate": 1.0024779126063412e-05,
+      "loss": 0.14,
+      "step": 400
+    },
+    {
+      "epoch": 1.9103773584905661,
+      "grad_norm": 1.1910700798034668,
+      "learning_rate": 9.658487462616795e-06,
+      "loss": 0.1756,
+      "step": 405
+    },
+    {
+      "epoch": 1.9339622641509435,
+      "grad_norm": 1.073564887046814,
+      "learning_rate": 9.295808280832154e-06,
+      "loss": 0.136,
+      "step": 410
+    },
+    {
+      "epoch": 1.9575471698113207,
+      "grad_norm": 1.272138237953186,
+      "learning_rate": 8.936986861776277e-06,
+      "loss": 0.1319,
+      "step": 415
+    },
+    {
+      "epoch": 1.9811320754716981,
+      "grad_norm": 1.0732004642486572,
+      "learning_rate": 8.582265877498911e-06,
+      "loss": 0.1186,
+      "step": 420
+    },
+    {
+      "epoch": 2.0047169811320753,
+      "grad_norm": 0.9422816634178162,
+      "learning_rate": 8.231885226912944e-06,
+      "loss": 0.1348,
+      "step": 425
+    },
+    {
+      "epoch": 2.0283018867924527,
+      "grad_norm": 1.181166410446167,
+      "learning_rate": 7.886081873549964e-06,
+      "loss": 0.1034,
+      "step": 430
+    },
+    {
+      "epoch": 2.05188679245283,
+      "grad_norm": 1.0801570415496826,
+      "learning_rate": 7.545089685301134e-06,
+      "loss": 0.0999,
+      "step": 435
+    },
+    {
+      "epoch": 2.0754716981132075,
+      "grad_norm": 0.8542254567146301,
+      "learning_rate": 7.209139276251559e-06,
+      "loss": 0.0973,
+      "step": 440
+    },
+    {
+      "epoch": 2.099056603773585,
+      "grad_norm": 1.0083342790603638,
+      "learning_rate": 6.878457850715293e-06,
+      "loss": 0.1098,
+      "step": 445
+    },
+    {
+      "epoch": 2.1226415094339623,
+      "grad_norm": 1.128522276878357,
+      "learning_rate": 6.553269049576317e-06,
+      "loss": 0.1081,
+      "step": 450
+    },
+    {
+      "epoch": 2.1462264150943398,
+      "grad_norm": 1.1913063526153564,
+      "learning_rate": 6.233792799039546e-06,
+      "loss": 0.0919,
+      "step": 455
+    },
+    {
+      "epoch": 2.169811320754717,
+      "grad_norm": 0.8760666251182556,
+      "learning_rate": 5.920245161894011e-06,
+      "loss": 0.0938,
+      "step": 460
+    },
+    {
+      "epoch": 2.1933962264150946,
+      "grad_norm": 0.9241344928741455,
+      "learning_rate": 5.612838191388936e-06,
+      "loss": 0.0978,
+      "step": 465
+    },
+    {
+      "epoch": 2.2169811320754715,
+      "grad_norm": 0.9444161057472229,
+      "learning_rate": 5.311779787821474e-06,
+      "loss": 0.1037,
+      "step": 470
+    },
+    {
+      "epoch": 2.240566037735849,
+      "grad_norm": 1.3713264465332031,
+      "learning_rate": 5.017273557933053e-06,
+      "loss": 0.0965,
+      "step": 475
+    },
+    {
+      "epoch": 2.2641509433962264,
+      "grad_norm": 0.9788404703140259,
+      "learning_rate": 4.729518677209523e-06,
+      "loss": 0.117,
+      "step": 480
+    },
+    {
+      "epoch": 2.2877358490566038,
+      "grad_norm": 0.8524981141090393,
+      "learning_rate": 4.4487097551781465e-06,
+      "loss": 0.0847,
+      "step": 485
+    },
+    {
+      "epoch": 2.311320754716981,
+      "grad_norm": 0.9140986204147339,
+      "learning_rate": 4.175036703792568e-06,
+      "loss": 0.0876,
+      "step": 490
+    },
+    {
+      "epoch": 2.3349056603773586,
+      "grad_norm": 0.6949209570884705,
+      "learning_rate": 3.908684608994748e-06,
+      "loss": 0.0824,
+      "step": 495
+    },
+    {
+      "epoch": 2.358490566037736,
+      "grad_norm": 0.8036820292472839,
+      "learning_rate": 3.6498336055407785e-06,
+      "loss": 0.0921,
+      "step": 500
+    },
+    {
+      "epoch": 2.3820754716981134,
+      "grad_norm": 0.8714132308959961,
+      "learning_rate": 3.398658755175183e-06,
+      "loss": 0.0935,
+      "step": 505
+    },
+    {
+      "epoch": 2.4056603773584904,
+      "grad_norm": 0.8123912811279297,
+      "learning_rate": 3.155329928236097e-06,
+      "loss": 0.0945,
+      "step": 510
+    },
+    {
+      "epoch": 2.4292452830188678,
+      "grad_norm": 0.9139612317085266,
+      "learning_rate": 2.9200116887714323e-06,
+      "loss": 0.0815,
+      "step": 515
+    },
+    {
+      "epoch": 2.452830188679245,
+      "grad_norm": 0.9900463223457336,
+      "learning_rate": 2.692863183243712e-06,
+      "loss": 0.0944,
+      "step": 520
+    },
+    {
+      "epoch": 2.4764150943396226,
+      "grad_norm": 0.9346297979354858,
+      "learning_rate": 2.474038032898786e-06,
+      "loss": 0.0873,
+      "step": 525
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 0.8317809700965881,
+      "learning_rate": 2.2636842298713355e-06,
+      "loss": 0.0996,
+      "step": 530
+    },
+    {
+      "epoch": 2.5235849056603774,
+      "grad_norm": 0.8242790102958679,
+      "learning_rate": 2.0619440370973077e-06,
+      "loss": 0.099,
+      "step": 535
+    },
+    {
+      "epoch": 2.547169811320755,
+      "grad_norm": 0.6480757594108582,
+      "learning_rate": 1.8689538921010485e-06,
+      "loss": 0.0815,
+      "step": 540
+    },
+    {
+      "epoch": 2.5707547169811322,
+      "grad_norm": 0.692046582698822,
+      "learning_rate": 1.6848443147221832e-06,
+      "loss": 0.0838,
+      "step": 545
+    },
+    {
+      "epoch": 2.5943396226415096,
+      "grad_norm": 0.9395256638526917,
+      "learning_rate": 1.5097398188446432e-06,
+      "loss": 0.0801,
+      "step": 550
+    },
+    {
+      "epoch": 2.617924528301887,
+      "grad_norm": 0.963858425617218,
+      "learning_rate": 1.3437588281875262e-06,
+      "loss": 0.085,
+      "step": 555
+    },
+    {
+      "epoch": 2.641509433962264,
+      "grad_norm": 0.7881783843040466,
+      "learning_rate": 1.1870135962147665e-06,
+      "loss": 0.0754,
+      "step": 560
+    },
+    {
+      "epoch": 2.6650943396226414,
+      "grad_norm": 0.8289976119995117,
+      "learning_rate": 1.0396101302177725e-06,
+      "loss": 0.088,
+      "step": 565
+    },
+    {
+      "epoch": 2.688679245283019,
+      "grad_norm": 0.8395482897758484,
+      "learning_rate": 9.016481196223563e-07,
+      "loss": 0.0829,
+      "step": 570
+    },
+    {
+      "epoch": 2.7122641509433962,
+      "grad_norm": 0.6427444815635681,
+      "learning_rate": 7.732208685684728e-07,
+      "loss": 0.0752,
+      "step": 575
+    },
+    {
+      "epoch": 2.7358490566037736,
+      "grad_norm": 0.7734426856040955,
+      "learning_rate": 6.544152328083153e-07,
+      "loss": 0.0807,
+      "step": 580
+    },
+    {
+      "epoch": 2.759433962264151,
+      "grad_norm": 0.7232455611228943,
+      "learning_rate": 5.453115609655285e-07,
+      "loss": 0.0762,
+      "step": 585
+    },
+    {
+      "epoch": 2.7830188679245285,
+      "grad_norm": 0.8498655557632446,
+      "learning_rate": 4.4598364019515835e-07,
+      "loss": 0.0878,
+      "step": 590
+    },
+    {
+      "epoch": 2.8066037735849054,
+      "grad_norm": 0.6923862099647522,
+      "learning_rate": 3.5649864628117535e-07,
+      "loss": 0.0852,
+      "step": 595
+    },
+    {
+      "epoch": 2.830188679245283,
+      "grad_norm": 0.76810622215271,
+      "learning_rate": 2.7691709820529043e-07,
+      "loss": 0.0733,
+      "step": 600
+    },
+    {
+      "epoch": 2.8537735849056602,
+      "grad_norm": 0.8895761966705322,
+      "learning_rate": 2.0729281721776682e-07,
+      "loss": 0.0784,
+      "step": 605
+    },
+    {
+      "epoch": 2.8773584905660377,
+      "grad_norm": 0.6226556897163391,
+      "learning_rate": 1.4767289043793986e-07,
+      "loss": 0.0713,
+      "step": 610
+    },
+    {
+      "epoch": 2.900943396226415,
+      "grad_norm": 0.7208622097969055,
+      "learning_rate": 9.809763900905878e-08,
+      "loss": 0.0691,
+      "step": 615
+    },
+    {
+      "epoch": 2.9245283018867925,
+      "grad_norm": 0.7647349834442139,
+      "learning_rate": 5.860059082896496e-08,
+      "loss": 0.0696,
+      "step": 620
+    },
+    {
+      "epoch": 2.94811320754717,
+      "grad_norm": 0.7331763505935669,
+      "learning_rate": 2.920845787507509e-08,
+      "loss": 0.0837,
+      "step": 625
+    },
+    {
+      "epoch": 2.9716981132075473,
+      "grad_norm": 0.7850632071495056,
+      "learning_rate": 9.941118138976912e-09,
+      "loss": 0.071,
+      "step": 630
+    },
+    {
+      "epoch": 2.9952830188679247,
+      "grad_norm": 0.6906620264053345,
+      "learning_rate": 8.116021828930009e-10,
+      "loss": 0.0843,
+      "step": 635
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 636,
+  "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": 8.24953730926379e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

125_128_e3_3e-5/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00033c960d81409c808104c084794b607d5f3605e731ba0a057dc85fe5944c9e
+size 8209

125_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)