Upload folder using huggingface_hub

13_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": [
+    "v_proj",
+    "k_proj",
+    "gate_proj",
+    "o_proj",
+    "down_proj",
+    "up_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

13_128_e3_3e-5/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0250bb072f127b55cd9b95d81d33d983b1cda181ed020473a29b530be3ec8b6
+size 791751704

13_128_e3_3e-5/latest

@@ -0,0 +1 @@
+global_step561

13_128_e3_3e-5/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37a326bb08a938d9a7a19ff22fba5e409f6d04f3ed09af8cfc63a6b8cdb75f87
+size 16389

13_128_e3_3e-5/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23d0c5b008f75e74e2bb33d46ecc0748966377a2a6e21be3d376e9611418ef88
+size 16389

13_128_e3_3e-5/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c96aab1ce05e509f2fd887cfe242e903beedbd36a1d8fc17f8d6cc639e73f2e2
+size 16389

13_128_e3_3e-5/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:51128346d6812dead0a499b979ced01b4b323450b433d376ed9438832f9d8ffe
+size 16389

13_128_e3_3e-5/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee22baace0b1959f6e514bdef1aea5c066ad85351a262df788bc6a1a63f7700c
+size 16389

13_128_e3_3e-5/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21daf33e3f9afc8abd91a7ec483e550fc3fec31289a5b2e596e6576fdd8b6505
+size 16389

13_128_e3_3e-5/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad29afc9c12b922b8d569ff7dcf7f2cf7f517e0aea5d203710e5ad924c22158b
+size 16389

13_128_e3_3e-5/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7d212c8066ada00e61382dec8715a3bfb4d0672a66a699f6b7746c9ad040cf27
+size 16389

13_128_e3_3e-5/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:56b97297f9f298b69ce6a2db2d75271b4143161dab5b26248a9c6d3f3bbda628
+size 1401

13_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
+  }
+}

13_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
+}

13_128_e3_3e-5/trainer_state.json

@@ -0,0 +1,818 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 561,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.02680965147453083,
+      "grad_norm": 0.8554138541221619,
+      "learning_rate": 4.137931034482758e-06,
+      "loss": 1.4474,
+      "step": 5
+    },
+    {
+      "epoch": 0.05361930294906166,
+      "grad_norm": 0.7120146751403809,
+      "learning_rate": 9.310344827586207e-06,
+      "loss": 1.4198,
+      "step": 10
+    },
+    {
+      "epoch": 0.08042895442359249,
+      "grad_norm": 0.585356593132019,
+      "learning_rate": 1.4482758620689657e-05,
+      "loss": 1.4144,
+      "step": 15
+    },
+    {
+      "epoch": 0.10723860589812333,
+      "grad_norm": 0.5165913701057434,
+      "learning_rate": 1.9655172413793102e-05,
+      "loss": 1.3441,
+      "step": 20
+    },
+    {
+      "epoch": 0.13404825737265416,
+      "grad_norm": 0.5417937636375427,
+      "learning_rate": 2.4827586206896553e-05,
+      "loss": 1.3148,
+      "step": 25
+    },
+    {
+      "epoch": 0.16085790884718498,
+      "grad_norm": 0.5062147378921509,
+      "learning_rate": 3e-05,
+      "loss": 1.2764,
+      "step": 30
+    },
+    {
+      "epoch": 0.1876675603217158,
+      "grad_norm": 0.5658809542655945,
+      "learning_rate": 2.999346198065638e-05,
+      "loss": 1.2304,
+      "step": 35
+    },
+    {
+      "epoch": 0.21447721179624665,
+      "grad_norm": 0.49390578269958496,
+      "learning_rate": 2.9973853622051763e-05,
+      "loss": 1.2031,
+      "step": 40
+    },
+    {
+      "epoch": 0.24128686327077747,
+      "grad_norm": 0.608232855796814,
+      "learning_rate": 2.9941192017496545e-05,
+      "loss": 1.1633,
+      "step": 45
+    },
+    {
+      "epoch": 0.2680965147453083,
+      "grad_norm": 0.6467882990837097,
+      "learning_rate": 2.989550563928436e-05,
+      "loss": 1.1463,
+      "step": 50
+    },
+    {
+      "epoch": 0.2949061662198391,
+      "grad_norm": 0.6315890550613403,
+      "learning_rate": 2.983683431387182e-05,
+      "loss": 1.0852,
+      "step": 55
+    },
+    {
+      "epoch": 0.32171581769436997,
+      "grad_norm": 0.6453232765197754,
+      "learning_rate": 2.9765229187160316e-05,
+      "loss": 1.024,
+      "step": 60
+    },
+    {
+      "epoch": 0.3485254691689008,
+      "grad_norm": 0.6697545647621155,
+      "learning_rate": 2.968075267991032e-05,
+      "loss": 1.0449,
+      "step": 65
+    },
+    {
+      "epoch": 0.3753351206434316,
+      "grad_norm": 0.7381000518798828,
+      "learning_rate": 2.958347843332696e-05,
+      "loss": 0.9543,
+      "step": 70
+    },
+    {
+      "epoch": 0.40214477211796246,
+      "grad_norm": 0.7937468886375427,
+      "learning_rate": 2.9473491244864347e-05,
+      "loss": 0.9233,
+      "step": 75
+    },
+    {
+      "epoch": 0.4289544235924933,
+      "grad_norm": 0.823359489440918,
+      "learning_rate": 2.9350886994304572e-05,
+      "loss": 0.9458,
+      "step": 80
+    },
+    {
+      "epoch": 0.45576407506702415,
+      "grad_norm": 0.9572771191596985,
+      "learning_rate": 2.9215772560175877e-05,
+      "loss": 0.9391,
+      "step": 85
+    },
+    {
+      "epoch": 0.48257372654155495,
+      "grad_norm": 0.8822752833366394,
+      "learning_rate": 2.906826572658278e-05,
+      "loss": 0.8435,
+      "step": 90
+    },
+    {
+      "epoch": 0.5093833780160858,
+      "grad_norm": 1.0733213424682617,
+      "learning_rate": 2.8908495080529468e-05,
+      "loss": 0.8169,
+      "step": 95
+    },
+    {
+      "epoch": 0.5361930294906166,
+      "grad_norm": 1.0479944944381714,
+      "learning_rate": 2.873659989982586e-05,
+      "loss": 0.7902,
+      "step": 100
+    },
+    {
+      "epoch": 0.5630026809651475,
+      "grad_norm": 0.9822242856025696,
+      "learning_rate": 2.8552730031674164e-05,
+      "loss": 0.7415,
+      "step": 105
+    },
+    {
+      "epoch": 0.5898123324396782,
+      "grad_norm": 1.0020015239715576,
+      "learning_rate": 2.835704576204167e-05,
+      "loss": 0.7208,
+      "step": 110
+    },
+    {
+      "epoch": 0.6166219839142091,
+      "grad_norm": 1.1209756135940552,
+      "learning_rate": 2.814971767593372e-05,
+      "loss": 0.6895,
+      "step": 115
+    },
+    {
+      "epoch": 0.6434316353887399,
+      "grad_norm": 1.3601207733154297,
+      "learning_rate": 2.7930926508688647e-05,
+      "loss": 0.6289,
+      "step": 120
+    },
+    {
+      "epoch": 0.6702412868632708,
+      "grad_norm": 1.0825207233428955,
+      "learning_rate": 2.770086298842426e-05,
+      "loss": 0.6226,
+      "step": 125
+    },
+    {
+      "epoch": 0.6970509383378016,
+      "grad_norm": 1.1475075483322144,
+      "learning_rate": 2.7459727669773344e-05,
+      "loss": 0.5403,
+      "step": 130
+    },
+    {
+      "epoch": 0.7238605898123325,
+      "grad_norm": 1.2114278078079224,
+      "learning_rate": 2.7207730759052925e-05,
+      "loss": 0.5699,
+      "step": 135
+    },
+    {
+      "epoch": 0.7506702412868632,
+      "grad_norm": 1.5098202228546143,
+      "learning_rate": 2.6945091931019916e-05,
+      "loss": 0.5721,
+      "step": 140
+    },
+    {
+      "epoch": 0.7774798927613941,
+      "grad_norm": 1.2076290845870972,
+      "learning_rate": 2.667204013737272e-05,
+      "loss": 0.5928,
+      "step": 145
+    },
+    {
+      "epoch": 0.8042895442359249,
+      "grad_norm": 1.1289753913879395,
+      "learning_rate": 2.638881340716583e-05,
+      "loss": 0.5965,
+      "step": 150
+    },
+    {
+      "epoch": 0.8310991957104558,
+      "grad_norm": 1.2910172939300537,
+      "learning_rate": 2.6095658639311348e-05,
+      "loss": 0.5215,
+      "step": 155
+    },
+    {
+      "epoch": 0.8579088471849866,
+      "grad_norm": 1.4134601354599,
+      "learning_rate": 2.5792831387348322e-05,
+      "loss": 0.5308,
+      "step": 160
+    },
+    {
+      "epoch": 0.8847184986595175,
+      "grad_norm": 1.5194172859191895,
+      "learning_rate": 2.548059563666757e-05,
+      "loss": 0.458,
+      "step": 165
+    },
+    {
+      "epoch": 0.9115281501340483,
+      "grad_norm": 1.233156681060791,
+      "learning_rate": 2.5159223574386117e-05,
+      "loss": 0.4562,
+      "step": 170
+    },
+    {
+      "epoch": 0.938337801608579,
+      "grad_norm": 1.2173125743865967,
+      "learning_rate": 2.4828995352071927e-05,
+      "loss": 0.4416,
+      "step": 175
+    },
+    {
+      "epoch": 0.9651474530831099,
+      "grad_norm": 1.353611946105957,
+      "learning_rate": 2.4490198841525708e-05,
+      "loss": 0.4287,
+      "step": 180
+    },
+    {
+      "epoch": 0.9919571045576407,
+      "grad_norm": 1.2814767360687256,
+      "learning_rate": 2.414312938383272e-05,
+      "loss": 0.427,
+      "step": 185
+    },
+    {
+      "epoch": 1.0160857908847185,
+      "grad_norm": 1.5797556638717651,
+      "learning_rate": 2.3788089531903372e-05,
+      "loss": 0.3875,
+      "step": 190
+    },
+    {
+      "epoch": 1.0428954423592494,
+      "grad_norm": 1.4800986051559448,
+      "learning_rate": 2.3425388786726944e-05,
+      "loss": 0.3386,
+      "step": 195
+    },
+    {
+      "epoch": 1.0697050938337802,
+      "grad_norm": 1.4225893020629883,
+      "learning_rate": 2.305534332756849e-05,
+      "loss": 0.375,
+      "step": 200
+    },
+    {
+      "epoch": 1.096514745308311,
+      "grad_norm": 1.3811910152435303,
+      "learning_rate": 2.2678275736344014e-05,
+      "loss": 0.3292,
+      "step": 205
+    },
+    {
+      "epoch": 1.123324396782842,
+      "grad_norm": 1.322371244430542,
+      "learning_rate": 2.229451471641422e-05,
+      "loss": 0.2958,
+      "step": 210
+    },
+    {
+      "epoch": 1.1501340482573728,
+      "grad_norm": 1.381089687347412,
+      "learning_rate": 2.190439480604199e-05,
+      "loss": 0.3101,
+      "step": 215
+    },
+    {
+      "epoch": 1.1769436997319036,
+      "grad_norm": 1.377687692642212,
+      "learning_rate": 2.1508256086763372e-05,
+      "loss": 0.3066,
+      "step": 220
+    },
+    {
+      "epoch": 1.2037533512064342,
+      "grad_norm": 1.2367618083953857,
+      "learning_rate": 2.1106443886926288e-05,
+      "loss": 0.2988,
+      "step": 225
+    },
+    {
+      "epoch": 1.230563002680965,
+      "grad_norm": 1.4667096138000488,
+      "learning_rate": 2.0699308480655397e-05,
+      "loss": 0.326,
+      "step": 230
+    },
+    {
+      "epoch": 1.257372654155496,
+      "grad_norm": 1.3385933637619019,
+      "learning_rate": 2.028720478250561e-05,
+      "loss": 0.2927,
+      "step": 235
+    },
+    {
+      "epoch": 1.2841823056300268,
+      "grad_norm": 1.414199709892273,
+      "learning_rate": 1.9870492038070255e-05,
+      "loss": 0.266,
+      "step": 240
+    },
+    {
+      "epoch": 1.3109919571045576,
+      "grad_norm": 1.4372038841247559,
+      "learning_rate": 1.944953351081385e-05,
+      "loss": 0.2571,
+      "step": 245
+    },
+    {
+      "epoch": 1.3378016085790885,
+      "grad_norm": 1.2661691904067993,
+      "learning_rate": 1.9024696165402272e-05,
+      "loss": 0.2368,
+      "step": 250
+    },
+    {
+      "epoch": 1.3646112600536193,
+      "grad_norm": 1.6702178716659546,
+      "learning_rate": 1.859635034780648e-05,
+      "loss": 0.2718,
+      "step": 255
+    },
+    {
+      "epoch": 1.3914209115281502,
+      "grad_norm": 1.5341639518737793,
+      "learning_rate": 1.8164869462458635e-05,
+      "loss": 0.2494,
+      "step": 260
+    },
+    {
+      "epoch": 1.418230563002681,
+      "grad_norm": 1.3405064344406128,
+      "learning_rate": 1.7730629646742038e-05,
+      "loss": 0.233,
+      "step": 265
+    },
+    {
+      "epoch": 1.4450402144772119,
+      "grad_norm": 1.1625229120254517,
+      "learning_rate": 1.729400944309869e-05,
+      "loss": 0.2115,
+      "step": 270
+    },
+    {
+      "epoch": 1.4718498659517425,
+      "grad_norm": 1.2672021389007568,
+      "learning_rate": 1.6855389469040217e-05,
+      "loss": 0.2323,
+      "step": 275
+    },
+    {
+      "epoch": 1.4986595174262733,
+      "grad_norm": 1.3835601806640625,
+      "learning_rate": 1.6415152085349934e-05,
+      "loss": 0.2219,
+      "step": 280
+    },
+    {
+      "epoch": 1.5254691689008042,
+      "grad_norm": 1.31817626953125,
+      "learning_rate": 1.597368106276523e-05,
+      "loss": 0.2131,
+      "step": 285
+    },
+    {
+      "epoch": 1.552278820375335,
+      "grad_norm": 1.1383699178695679,
+      "learning_rate": 1.553136124743081e-05,
+      "loss": 0.1868,
+      "step": 290
+    },
+    {
+      "epoch": 1.579088471849866,
+      "grad_norm": 1.1617040634155273,
+      "learning_rate": 1.50885782254145e-05,
+      "loss": 0.2112,
+      "step": 295
+    },
+    {
+      "epoch": 1.6058981233243967,
+      "grad_norm": 1.411847710609436,
+      "learning_rate": 1.464571798657803e-05,
+      "loss": 0.1984,
+      "step": 300
+    },
+    {
+      "epoch": 1.6327077747989276,
+      "grad_norm": 1.5382616519927979,
+      "learning_rate": 1.4203166588095816e-05,
+      "loss": 0.1932,
+      "step": 305
+    },
+    {
+      "epoch": 1.6595174262734584,
+      "grad_norm": 1.485434889793396,
+      "learning_rate": 1.3761309817915017e-05,
+      "loss": 0.189,
+      "step": 310
+    },
+    {
+      "epoch": 1.6863270777479893,
+      "grad_norm": 1.3767578601837158,
+      "learning_rate": 1.3320532858450382e-05,
+      "loss": 0.1727,
+      "step": 315
+    },
+    {
+      "epoch": 1.7131367292225201,
+      "grad_norm": 1.0452196598052979,
+      "learning_rate": 1.2881219950806875e-05,
+      "loss": 0.1938,
+      "step": 320
+    },
+    {
+      "epoch": 1.739946380697051,
+      "grad_norm": 1.477149248123169,
+      "learning_rate": 1.2443754059822901e-05,
+      "loss": 0.1561,
+      "step": 325
+    },
+    {
+      "epoch": 1.7667560321715818,
+      "grad_norm": 1.2107077836990356,
+      "learning_rate": 1.2008516540226115e-05,
+      "loss": 0.1692,
+      "step": 330
+    },
+    {
+      "epoch": 1.7935656836461127,
+      "grad_norm": 1.325714111328125,
+      "learning_rate": 1.1575886804192816e-05,
+      "loss": 0.1686,
+      "step": 335
+    },
+    {
+      "epoch": 1.8203753351206435,
+      "grad_norm": 1.3974460363388062,
+      "learning_rate": 1.1146241990600713e-05,
+      "loss": 0.1486,
+      "step": 340
+    },
+    {
+      "epoch": 1.8471849865951744,
+      "grad_norm": 1.2481837272644043,
+      "learning_rate": 1.0719956636263424e-05,
+      "loss": 0.1374,
+      "step": 345
+    },
+    {
+      "epoch": 1.8739946380697052,
+      "grad_norm": 1.3408443927764893,
+      "learning_rate": 1.0297402349433286e-05,
+      "loss": 0.1408,
+      "step": 350
+    },
+    {
+      "epoch": 1.900804289544236,
+      "grad_norm": 1.1914774179458618,
+      "learning_rate": 9.878947485857104e-06,
+      "loss": 0.1504,
+      "step": 355
+    },
+    {
+      "epoch": 1.9276139410187667,
+      "grad_norm": 1.3521840572357178,
+      "learning_rate": 9.464956827667216e-06,
+      "loss": 0.1677,
+      "step": 360
+    },
+    {
+      "epoch": 1.9544235924932976,
+      "grad_norm": 1.0081520080566406,
+      "learning_rate": 9.05579126538779e-06,
+      "loss": 0.1458,
+      "step": 365
+    },
+    {
+      "epoch": 1.9812332439678284,
+      "grad_norm": 1.011947512626648,
+      "learning_rate": 8.651807483333627e-06,
+      "loss": 0.1241,
+      "step": 370
+    },
+    {
+      "epoch": 2.005361930294906,
+      "grad_norm": 1.0631119012832642,
+      "learning_rate": 8.253357648675594e-06,
+      "loss": 0.1198,
+      "step": 375
+    },
+    {
+      "epoch": 2.032171581769437,
+      "grad_norm": 1.3207858800888062,
+      "learning_rate": 7.860789104443897e-06,
+      "loss": 0.0874,
+      "step": 380
+    },
+    {
+      "epoch": 2.058981233243968,
+      "grad_norm": 1.0557644367218018,
+      "learning_rate": 7.474444066736647e-06,
+      "loss": 0.112,
+      "step": 385
+    },
+    {
+      "epoch": 2.0857908847184987,
+      "grad_norm": 0.9301002025604248,
+      "learning_rate": 7.094659326397818e-06,
+      "loss": 0.097,
+      "step": 390
+    },
+    {
+      "epoch": 2.1126005361930296,
+      "grad_norm": 0.8232920169830322,
+      "learning_rate": 6.721765955424589e-06,
+      "loss": 0.1059,
+      "step": 395
+    },
+    {
+      "epoch": 2.1394101876675604,
+      "grad_norm": 0.8654112219810486,
+      "learning_rate": 6.35608901835995e-06,
+      "loss": 0.0857,
+      "step": 400
+    },
+    {
+      "epoch": 2.1662198391420913,
+      "grad_norm": 1.091780424118042,
+      "learning_rate": 5.99794728892232e-06,
+      "loss": 0.1003,
+      "step": 405
+    },
+    {
+      "epoch": 2.193029490616622,
+      "grad_norm": 0.8873447775840759,
+      "learning_rate": 5.647652972118998e-06,
+      "loss": 0.0969,
+      "step": 410
+    },
+    {
+      "epoch": 2.219839142091153,
+      "grad_norm": 0.9298948049545288,
+      "learning_rate": 5.305511432085885e-06,
+      "loss": 0.0928,
+      "step": 415
+    },
+    {
+      "epoch": 2.246648793565684,
+      "grad_norm": 0.9384891390800476,
+      "learning_rate": 4.9718209258905865e-06,
+      "loss": 0.0998,
+      "step": 420
+    },
+    {
+      "epoch": 2.2734584450402147,
+      "grad_norm": 0.9399723410606384,
+      "learning_rate": 4.646872343530999e-06,
+      "loss": 0.0952,
+      "step": 425
+    },
+    {
+      "epoch": 2.3002680965147455,
+      "grad_norm": 0.9280492663383484,
+      "learning_rate": 4.330948954356076e-06,
+      "loss": 0.0913,
+      "step": 430
+    },
+    {
+      "epoch": 2.327077747989276,
+      "grad_norm": 0.9419163465499878,
+      "learning_rate": 4.024326160129761e-06,
+      "loss": 0.0957,
+      "step": 435
+    },
+    {
+      "epoch": 2.353887399463807,
+      "grad_norm": 0.9238213300704956,
+      "learning_rate": 3.7272712549533568e-06,
+      "loss": 0.0834,
+      "step": 440
+    },
+    {
+      "epoch": 2.3806970509383376,
+      "grad_norm": 0.814352810382843,
+      "learning_rate": 3.440043192255693e-06,
+      "loss": 0.0884,
+      "step": 445
+    },
+    {
+      "epoch": 2.4075067024128685,
+      "grad_norm": 0.7930625677108765,
+      "learning_rate": 3.162892359054098e-06,
+      "loss": 0.0875,
+      "step": 450
+    },
+    {
+      "epoch": 2.4343163538873993,
+      "grad_norm": 0.8555788993835449,
+      "learning_rate": 2.8960603576830408e-06,
+      "loss": 0.0907,
+      "step": 455
+    },
+    {
+      "epoch": 2.46112600536193,
+      "grad_norm": 0.8803578019142151,
+      "learning_rate": 2.63977979518072e-06,
+      "loss": 0.0994,
+      "step": 460
+    },
+    {
+      "epoch": 2.487935656836461,
+      "grad_norm": 0.8803998827934265,
+      "learning_rate": 2.394274080517143e-06,
+      "loss": 0.0791,
+      "step": 465
+    },
+    {
+      "epoch": 2.514745308310992,
+      "grad_norm": 0.9126463532447815,
+      "learning_rate": 2.1597572298405e-06,
+      "loss": 0.0699,
+      "step": 470
+    },
+    {
+      "epoch": 2.5415549597855227,
+      "grad_norm": 0.695434033870697,
+      "learning_rate": 1.936433679911604e-06,
+      "loss": 0.0769,
+      "step": 475
+    },
+    {
+      "epoch": 2.5683646112600536,
+      "grad_norm": 0.7444580793380737,
+      "learning_rate": 1.724498109889039e-06,
+      "loss": 0.0745,
+      "step": 480
+    },
+    {
+      "epoch": 2.5951742627345844,
+      "grad_norm": 0.8400862812995911,
+      "learning_rate": 1.5241352716203173e-06,
+      "loss": 0.0777,
+      "step": 485
+    },
+    {
+      "epoch": 2.6219839142091153,
+      "grad_norm": 0.8520445823669434,
+      "learning_rate": 1.3355198285870935e-06,
+      "loss": 0.0838,
+      "step": 490
+    },
+    {
+      "epoch": 2.648793565683646,
+      "grad_norm": 0.7402787804603577,
+      "learning_rate": 1.1588162036447003e-06,
+      "loss": 0.0652,
+      "step": 495
+    },
+    {
+      "epoch": 2.675603217158177,
+      "grad_norm": 0.8451915383338928,
+      "learning_rate": 9.941784356888717e-07,
+      "loss": 0.0757,
+      "step": 500
+    },
+    {
+      "epoch": 2.702412868632708,
+      "grad_norm": 0.8868169784545898,
+      "learning_rate": 8.417500453744864e-07,
+      "loss": 0.0871,
+      "step": 505
+    },
+    {
+      "epoch": 2.7292225201072386,
+      "grad_norm": 0.7002440094947815,
+      "learning_rate": 7.016639100034627e-07,
+      "loss": 0.0712,
+      "step": 510
+    },
+    {
+      "epoch": 2.7560321715817695,
+      "grad_norm": 0.9022250771522522,
+      "learning_rate": 5.740421476908391e-07,
+      "loss": 0.0778,
+      "step": 515
+    },
+    {
+      "epoch": 2.7828418230563003,
+      "grad_norm": 0.9231407642364502,
+      "learning_rate": 4.589960109100444e-07,
+      "loss": 0.0694,
+      "step": 520
+    },
+    {
+      "epoch": 2.809651474530831,
+      "grad_norm": 0.7728066444396973,
+      "learning_rate": 3.566257895101005e-07,
+      "loss": 0.0722,
+      "step": 525
+    },
+    {
+      "epoch": 2.836461126005362,
+      "grad_norm": 0.8146858215332031,
+      "learning_rate": 2.670207232893684e-07,
+      "loss": 0.072,
+      "step": 530
+    },
+    {
+      "epoch": 2.863270777479893,
+      "grad_norm": 0.6987660527229309,
+      "learning_rate": 1.9025892420201807e-07,
+      "loss": 0.0679,
+      "step": 535
+    },
+    {
+      "epoch": 2.8900804289544237,
+      "grad_norm": 1.019551396369934,
+      "learning_rate": 1.264073082650169e-07,
+      "loss": 0.0759,
+      "step": 540
+    },
+    {
+      "epoch": 2.9168900804289546,
+      "grad_norm": 0.7102984189987183,
+      "learning_rate": 7.55215372250484e-08,
+      "loss": 0.0692,
+      "step": 545
+    },
+    {
+      "epoch": 2.943699731903485,
+      "grad_norm": 0.7476786971092224,
+      "learning_rate": 3.764597003616421e-08,
+      "loss": 0.0775,
+      "step": 550
+    },
+    {
+      "epoch": 2.9705093833780163,
+      "grad_norm": 0.7322636842727661,
+      "learning_rate": 1.2813624190484707e-08,
+      "loss": 0.0755,
+      "step": 555
+    },
+    {
+      "epoch": 2.9973190348525467,
+      "grad_norm": 0.7615118026733398,
+      "learning_rate": 1.0461469356776165e-09,
+      "loss": 0.0823,
+      "step": 560
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 561,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 2000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 7.428659857441423e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

13_128_e3_3e-5/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8f2e291636ef183aa8884d7f78ed680f435eabe7b04969923f22f5787dcbadd
+size 8209

13_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)