Training in progress, step 100, checkpoint

checkpoint-100/config.json

@@ -0,0 +1,31 @@
+{
+  "architectures": [
+    "GPTNeoXForCausalLM"
+  ],
+  "attention_bias": true,
+  "attention_dropout": 0.0,
+  "bos_token_id": 0,
+  "classifier_dropout": 0.1,
+  "eos_token_id": 0,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.0,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 2048,
+  "model_type": "gpt_neox",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "partial_rotary_factor": 0.25,
+  "rope_scaling": null,
+  "rope_theta": 10000,
+  "rotary_emb_base": 10000,
+  "rotary_pct": 0.25,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.50.2",
+  "use_cache": true,
+  "use_parallel_residual": true,
+  "vocab_size": 50304
+}

checkpoint-100/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 0,
+  "eos_token_id": 0,
+  "transformers_version": "4.50.2"
+}

checkpoint-100/global_step100/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e641280eb8054f489749c724bf9460f7141b6c4e480a8c66f6c9bbf2c2595eb
+size 810750072

checkpoint-100/global_step100/zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42c329e2c752996f937fc0931fbb1579fd1180b86b85c09ba9849c6f6d4f3888
+size 1216012660

checkpoint-100/global_step100/zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a672cde24f064fb3f0a7275d01ea3929f16908fd3282825c368c32f0255bb25
+size 1216014068

checkpoint-100/global_step100/zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82fb79f97f233b414455a1727fbd834d1dec9d62101acea1a87777fc98449e61
+size 1216014132

checkpoint-100/global_step100/zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9200c39e02038c09e9adb1d711d8751fd8acb05aef61232b0ea95b1c6b48619e
+size 1216012724

checkpoint-100/latest

@@ -0,0 +1 @@
+global_step100

checkpoint-100/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac971311f9b703711e45b34dc62e9b5e3945bef27cb42dc1103fe14ef73f0426
+size 810701896

checkpoint-100/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36642765f9acf7ac8a4e1a3bfa5ea9c22a142cdfbfeeb96b983e197f754a4b1e
+size 14960

checkpoint-100/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd9f71d8de3875b8d2f115e69a265a76fb6ee4b01473a67358cddd05b7ac1ba1
+size 14960

checkpoint-100/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:86fef6e746cbac928ab2e67eac71866b327c2f765173f28a46ee21ddcc06220f
+size 14960

checkpoint-100/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a94f5a7d4c7e687cef690df30e6e15ad66167456ca9f35d50069b319c7fe11ec
+size 14960

checkpoint-100/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fbe02124a5bd070b2f718c06054742287515af164f6576de7eec1568bdfa7b28
+size 1256

checkpoint-100/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "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": "<|endoftext|>",
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-100/tokenizer_config.json

@@ -0,0 +1,215 @@
+{
+  "add_bos_token": false,
+  "add_eos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<|padding|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "50254": {
+      "content": "                        ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50255": {
+      "content": "                       ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50256": {
+      "content": "                      ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50257": {
+      "content": "                     ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50258": {
+      "content": "                    ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50259": {
+      "content": "                   ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50260": {
+      "content": "                  ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50261": {
+      "content": "                 ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50262": {
+      "content": "                ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50263": {
+      "content": "               ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50264": {
+      "content": "              ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50265": {
+      "content": "             ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50266": {
+      "content": "            ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50267": {
+      "content": "           ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50268": {
+      "content": "          ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50269": {
+      "content": "         ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50270": {
+      "content": "        ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50271": {
+      "content": "       ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50272": {
+      "content": "      ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50273": {
+      "content": "     ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50274": {
+      "content": "    ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50275": {
+      "content": "   ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50276": {
+      "content": "  ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "extra_special_tokens": {},
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<|endoftext|>",
+  "tokenizer_class": "GPTNeoXTokenizer",
+  "unk_token": "<|endoftext|>"
+}

checkpoint-100/trainer_state.json

@@ -0,0 +1,734 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.006993006993006993,
+  "eval_steps": 500,
+  "global_step": 100,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 6.993006993006993e-05,
+      "grad_norm": 5.159224033355713,
+      "learning_rate": 1.048951048951049e-06,
+      "loss": 11.0749,
+      "step": 1
+    },
+    {
+      "epoch": 0.00013986013986013986,
+      "grad_norm": 5.171240329742432,
+      "learning_rate": 2.097902097902098e-06,
+      "loss": 11.076,
+      "step": 2
+    },
+    {
+      "epoch": 0.0002097902097902098,
+      "grad_norm": 5.199888229370117,
+      "learning_rate": 3.146853146853147e-06,
+      "loss": 11.0669,
+      "step": 3
+    },
+    {
+      "epoch": 0.0002797202797202797,
+      "grad_norm": 5.08656120300293,
+      "learning_rate": 4.195804195804196e-06,
+      "loss": 11.0237,
+      "step": 4
+    },
+    {
+      "epoch": 0.00034965034965034965,
+      "grad_norm": 5.065122127532959,
+      "learning_rate": 5.244755244755245e-06,
+      "loss": 10.8949,
+      "step": 5
+    },
+    {
+      "epoch": 0.0004195804195804196,
+      "grad_norm": 4.298279762268066,
+      "learning_rate": 6.293706293706294e-06,
+      "loss": 10.675,
+      "step": 6
+    },
+    {
+      "epoch": 0.0004895104895104895,
+      "grad_norm": 3.8743011951446533,
+      "learning_rate": 7.342657342657342e-06,
+      "loss": 10.4729,
+      "step": 7
+    },
+    {
+      "epoch": 0.0005594405594405594,
+      "grad_norm": 3.7294981479644775,
+      "learning_rate": 8.391608391608391e-06,
+      "loss": 10.3221,
+      "step": 8
+    },
+    {
+      "epoch": 0.0006293706293706294,
+      "grad_norm": 3.3471932411193848,
+      "learning_rate": 9.44055944055944e-06,
+      "loss": 10.1544,
+      "step": 9
+    },
+    {
+      "epoch": 0.0006993006993006993,
+      "grad_norm": 3.088207721710205,
+      "learning_rate": 1.048951048951049e-05,
+      "loss": 10.0823,
+      "step": 10
+    },
+    {
+      "epoch": 0.0007692307692307692,
+      "grad_norm": 2.713122606277466,
+      "learning_rate": 1.1538461538461538e-05,
+      "loss": 9.9324,
+      "step": 11
+    },
+    {
+      "epoch": 0.0008391608391608392,
+      "grad_norm": 2.5934906005859375,
+      "learning_rate": 1.2587412587412587e-05,
+      "loss": 9.8238,
+      "step": 12
+    },
+    {
+      "epoch": 0.0009090909090909091,
+      "grad_norm": 2.258924722671509,
+      "learning_rate": 1.3636363636363635e-05,
+      "loss": 9.7534,
+      "step": 13
+    },
+    {
+      "epoch": 0.000979020979020979,
+      "grad_norm": 2.0431277751922607,
+      "learning_rate": 1.4685314685314684e-05,
+      "loss": 9.683,
+      "step": 14
+    },
+    {
+      "epoch": 0.001048951048951049,
+      "grad_norm": 2.0174779891967773,
+      "learning_rate": 1.5734265734265734e-05,
+      "loss": 9.5835,
+      "step": 15
+    },
+    {
+      "epoch": 0.0011188811188811189,
+      "grad_norm": 1.9392926692962646,
+      "learning_rate": 1.6783216783216783e-05,
+      "loss": 9.566,
+      "step": 16
+    },
+    {
+      "epoch": 0.0011888111888111888,
+      "grad_norm": 1.841917634010315,
+      "learning_rate": 1.783216783216783e-05,
+      "loss": 9.4862,
+      "step": 17
+    },
+    {
+      "epoch": 0.0012587412587412587,
+      "grad_norm": 1.6751583814620972,
+      "learning_rate": 1.888111888111888e-05,
+      "loss": 9.4613,
+      "step": 18
+    },
+    {
+      "epoch": 0.0013286713286713287,
+      "grad_norm": 1.701292634010315,
+      "learning_rate": 1.9930069930069928e-05,
+      "loss": 9.3861,
+      "step": 19
+    },
+    {
+      "epoch": 0.0013986013986013986,
+      "grad_norm": 1.7422462701797485,
+      "learning_rate": 2.097902097902098e-05,
+      "loss": 9.3607,
+      "step": 20
+    },
+    {
+      "epoch": 0.0014685314685314685,
+      "grad_norm": 1.7331982851028442,
+      "learning_rate": 2.2027972027972026e-05,
+      "loss": 9.3106,
+      "step": 21
+    },
+    {
+      "epoch": 0.0015384615384615385,
+      "grad_norm": 1.8083829879760742,
+      "learning_rate": 2.3076923076923076e-05,
+      "loss": 9.252,
+      "step": 22
+    },
+    {
+      "epoch": 0.0016083916083916084,
+      "grad_norm": 1.7314563989639282,
+      "learning_rate": 2.412587412587412e-05,
+      "loss": 9.2325,
+      "step": 23
+    },
+    {
+      "epoch": 0.0016783216783216783,
+      "grad_norm": 1.5793718099594116,
+      "learning_rate": 2.5174825174825174e-05,
+      "loss": 9.2489,
+      "step": 24
+    },
+    {
+      "epoch": 0.0017482517482517483,
+      "grad_norm": 1.5183552503585815,
+      "learning_rate": 2.622377622377622e-05,
+      "loss": 9.1821,
+      "step": 25
+    },
+    {
+      "epoch": 0.0018181818181818182,
+      "grad_norm": 1.5751609802246094,
+      "learning_rate": 2.727272727272727e-05,
+      "loss": 9.1217,
+      "step": 26
+    },
+    {
+      "epoch": 0.0018881118881118881,
+      "grad_norm": 1.541367530822754,
+      "learning_rate": 2.832167832167832e-05,
+      "loss": 9.1186,
+      "step": 27
+    },
+    {
+      "epoch": 0.001958041958041958,
+      "grad_norm": 1.5820801258087158,
+      "learning_rate": 2.937062937062937e-05,
+      "loss": 9.105,
+      "step": 28
+    },
+    {
+      "epoch": 0.002027972027972028,
+      "grad_norm": 1.483319640159607,
+      "learning_rate": 3.0419580419580414e-05,
+      "loss": 9.0772,
+      "step": 29
+    },
+    {
+      "epoch": 0.002097902097902098,
+      "grad_norm": 1.5269867181777954,
+      "learning_rate": 3.146853146853147e-05,
+      "loss": 9.0185,
+      "step": 30
+    },
+    {
+      "epoch": 0.002167832167832168,
+      "grad_norm": 1.4273761510849,
+      "learning_rate": 3.251748251748251e-05,
+      "loss": 9.0013,
+      "step": 31
+    },
+    {
+      "epoch": 0.0022377622377622378,
+      "grad_norm": 1.4377866983413696,
+      "learning_rate": 3.3566433566433566e-05,
+      "loss": 8.9161,
+      "step": 32
+    },
+    {
+      "epoch": 0.002307692307692308,
+      "grad_norm": 1.369158148765564,
+      "learning_rate": 3.461538461538461e-05,
+      "loss": 8.917,
+      "step": 33
+    },
+    {
+      "epoch": 0.0023776223776223776,
+      "grad_norm": 1.328121542930603,
+      "learning_rate": 3.566433566433566e-05,
+      "loss": 8.8484,
+      "step": 34
+    },
+    {
+      "epoch": 0.0024475524475524478,
+      "grad_norm": 1.289856195449829,
+      "learning_rate": 3.671328671328671e-05,
+      "loss": 8.8637,
+      "step": 35
+    },
+    {
+      "epoch": 0.0025174825174825175,
+      "grad_norm": 1.2946962118148804,
+      "learning_rate": 3.776223776223776e-05,
+      "loss": 8.8045,
+      "step": 36
+    },
+    {
+      "epoch": 0.0025874125874125876,
+      "grad_norm": 1.3396406173706055,
+      "learning_rate": 3.881118881118881e-05,
+      "loss": 8.7569,
+      "step": 37
+    },
+    {
+      "epoch": 0.0026573426573426573,
+      "grad_norm": 1.3043384552001953,
+      "learning_rate": 3.9860139860139855e-05,
+      "loss": 8.7184,
+      "step": 38
+    },
+    {
+      "epoch": 0.0027272727272727275,
+      "grad_norm": 1.2441431283950806,
+      "learning_rate": 4.09090909090909e-05,
+      "loss": 8.66,
+      "step": 39
+    },
+    {
+      "epoch": 0.002797202797202797,
+      "grad_norm": 1.1379547119140625,
+      "learning_rate": 4.195804195804196e-05,
+      "loss": 8.6217,
+      "step": 40
+    },
+    {
+      "epoch": 0.0028671328671328673,
+      "grad_norm": 1.1792885065078735,
+      "learning_rate": 4.300699300699301e-05,
+      "loss": 8.5006,
+      "step": 41
+    },
+    {
+      "epoch": 0.002937062937062937,
+      "grad_norm": 1.183509349822998,
+      "learning_rate": 4.405594405594405e-05,
+      "loss": 8.5037,
+      "step": 42
+    },
+    {
+      "epoch": 0.003006993006993007,
+      "grad_norm": 1.1481878757476807,
+      "learning_rate": 4.51048951048951e-05,
+      "loss": 8.5114,
+      "step": 43
+    },
+    {
+      "epoch": 0.003076923076923077,
+      "grad_norm": 1.1832213401794434,
+      "learning_rate": 4.615384615384615e-05,
+      "loss": 8.4117,
+      "step": 44
+    },
+    {
+      "epoch": 0.003146853146853147,
+      "grad_norm": 1.0428388118743896,
+      "learning_rate": 4.72027972027972e-05,
+      "loss": 8.4378,
+      "step": 45
+    },
+    {
+      "epoch": 0.0032167832167832168,
+      "grad_norm": 1.1736469268798828,
+      "learning_rate": 4.825174825174824e-05,
+      "loss": 8.2981,
+      "step": 46
+    },
+    {
+      "epoch": 0.003286713286713287,
+      "grad_norm": 1.0533231496810913,
+      "learning_rate": 4.930069930069929e-05,
+      "loss": 8.3406,
+      "step": 47
+    },
+    {
+      "epoch": 0.0033566433566433566,
+      "grad_norm": 1.1464074850082397,
+      "learning_rate": 5.034965034965035e-05,
+      "loss": 8.2388,
+      "step": 48
+    },
+    {
+      "epoch": 0.003426573426573427,
+      "grad_norm": 1.224696159362793,
+      "learning_rate": 5.1398601398601395e-05,
+      "loss": 8.208,
+      "step": 49
+    },
+    {
+      "epoch": 0.0034965034965034965,
+      "grad_norm": 1.1058298349380493,
+      "learning_rate": 5.244755244755244e-05,
+      "loss": 8.192,
+      "step": 50
+    },
+    {
+      "epoch": 0.0035664335664335666,
+      "grad_norm": 1.0586011409759521,
+      "learning_rate": 5.3496503496503493e-05,
+      "loss": 8.1369,
+      "step": 51
+    },
+    {
+      "epoch": 0.0036363636363636364,
+      "grad_norm": 1.0969116687774658,
+      "learning_rate": 5.454545454545454e-05,
+      "loss": 8.0515,
+      "step": 52
+    },
+    {
+      "epoch": 0.0037062937062937065,
+      "grad_norm": 0.9527297616004944,
+      "learning_rate": 5.559440559440559e-05,
+      "loss": 8.0319,
+      "step": 53
+    },
+    {
+      "epoch": 0.0037762237762237762,
+      "grad_norm": 1.0115875005722046,
+      "learning_rate": 5.664335664335664e-05,
+      "loss": 7.9882,
+      "step": 54
+    },
+    {
+      "epoch": 0.0038461538461538464,
+      "grad_norm": 1.044940710067749,
+      "learning_rate": 5.769230769230769e-05,
+      "loss": 7.9776,
+      "step": 55
+    },
+    {
+      "epoch": 0.003916083916083916,
+      "grad_norm": 0.9222269058227539,
+      "learning_rate": 5.874125874125874e-05,
+      "loss": 7.9327,
+      "step": 56
+    },
+    {
+      "epoch": 0.003986013986013986,
+      "grad_norm": 0.9342182278633118,
+      "learning_rate": 5.979020979020978e-05,
+      "loss": 7.853,
+      "step": 57
+    },
+    {
+      "epoch": 0.004055944055944056,
+      "grad_norm": 0.9146339297294617,
+      "learning_rate": 6.083916083916083e-05,
+      "loss": 7.8091,
+      "step": 58
+    },
+    {
+      "epoch": 0.004125874125874126,
+      "grad_norm": 0.8632137775421143,
+      "learning_rate": 6.188811188811188e-05,
+      "loss": 7.7744,
+      "step": 59
+    },
+    {
+      "epoch": 0.004195804195804196,
+      "grad_norm": 1.001086711883545,
+      "learning_rate": 6.293706293706293e-05,
+      "loss": 7.7623,
+      "step": 60
+    },
+    {
+      "epoch": 0.0042657342657342655,
+      "grad_norm": 0.9110626578330994,
+      "learning_rate": 6.398601398601397e-05,
+      "loss": 7.7356,
+      "step": 61
+    },
+    {
+      "epoch": 0.004335664335664336,
+      "grad_norm": 0.7626254558563232,
+      "learning_rate": 6.503496503496503e-05,
+      "loss": 7.6801,
+      "step": 62
+    },
+    {
+      "epoch": 0.004405594405594406,
+      "grad_norm": 1.0254472494125366,
+      "learning_rate": 6.608391608391608e-05,
+      "loss": 7.6111,
+      "step": 63
+    },
+    {
+      "epoch": 0.0044755244755244755,
+      "grad_norm": 0.9034542441368103,
+      "learning_rate": 6.713286713286713e-05,
+      "loss": 7.603,
+      "step": 64
+    },
+    {
+      "epoch": 0.004545454545454545,
+      "grad_norm": 0.685672402381897,
+      "learning_rate": 6.818181818181817e-05,
+      "loss": 7.5614,
+      "step": 65
+    },
+    {
+      "epoch": 0.004615384615384616,
+      "grad_norm": 0.9960275292396545,
+      "learning_rate": 6.923076923076922e-05,
+      "loss": 7.5537,
+      "step": 66
+    },
+    {
+      "epoch": 0.0046853146853146855,
+      "grad_norm": 0.6702042818069458,
+      "learning_rate": 7.027972027972028e-05,
+      "loss": 7.4794,
+      "step": 67
+    },
+    {
+      "epoch": 0.004755244755244755,
+      "grad_norm": 0.8843338489532471,
+      "learning_rate": 7.132867132867132e-05,
+      "loss": 7.4486,
+      "step": 68
+    },
+    {
+      "epoch": 0.004825174825174825,
+      "grad_norm": 0.7096754908561707,
+      "learning_rate": 7.237762237762237e-05,
+      "loss": 7.4253,
+      "step": 69
+    },
+    {
+      "epoch": 0.0048951048951048955,
+      "grad_norm": 0.8532392382621765,
+      "learning_rate": 7.342657342657342e-05,
+      "loss": 7.383,
+      "step": 70
+    },
+    {
+      "epoch": 0.004965034965034965,
+      "grad_norm": 0.7407330870628357,
+      "learning_rate": 7.447552447552447e-05,
+      "loss": 7.4059,
+      "step": 71
+    },
+    {
+      "epoch": 0.005034965034965035,
+      "grad_norm": 0.7601925730705261,
+      "learning_rate": 7.552447552447553e-05,
+      "loss": 7.3522,
+      "step": 72
+    },
+    {
+      "epoch": 0.005104895104895105,
+      "grad_norm": 0.6095029711723328,
+      "learning_rate": 7.657342657342657e-05,
+      "loss": 7.3442,
+      "step": 73
+    },
+    {
+      "epoch": 0.005174825174825175,
+      "grad_norm": 0.8200113773345947,
+      "learning_rate": 7.762237762237762e-05,
+      "loss": 7.3272,
+      "step": 74
+    },
+    {
+      "epoch": 0.005244755244755245,
+      "grad_norm": 0.9570252299308777,
+      "learning_rate": 7.867132867132867e-05,
+      "loss": 7.2258,
+      "step": 75
+    },
+    {
+      "epoch": 0.005314685314685315,
+      "grad_norm": 1.005595088005066,
+      "learning_rate": 7.972027972027971e-05,
+      "loss": 7.2552,
+      "step": 76
+    },
+    {
+      "epoch": 0.005384615384615384,
+      "grad_norm": 0.7686155438423157,
+      "learning_rate": 8.076923076923076e-05,
+      "loss": 7.2615,
+      "step": 77
+    },
+    {
+      "epoch": 0.005454545454545455,
+      "grad_norm": 0.6844556927680969,
+      "learning_rate": 8.18181818181818e-05,
+      "loss": 7.1683,
+      "step": 78
+    },
+    {
+      "epoch": 0.005524475524475525,
+      "grad_norm": 0.665515661239624,
+      "learning_rate": 8.286713286713286e-05,
+      "loss": 7.2015,
+      "step": 79
+    },
+    {
+      "epoch": 0.005594405594405594,
+      "grad_norm": 0.7764729261398315,
+      "learning_rate": 8.391608391608392e-05,
+      "loss": 7.1474,
+      "step": 80
+    },
+    {
+      "epoch": 0.005664335664335664,
+      "grad_norm": 0.6877023577690125,
+      "learning_rate": 8.496503496503496e-05,
+      "loss": 7.1705,
+      "step": 81
+    },
+    {
+      "epoch": 0.005734265734265735,
+      "grad_norm": 0.678676426410675,
+      "learning_rate": 8.601398601398601e-05,
+      "loss": 7.1529,
+      "step": 82
+    },
+    {
+      "epoch": 0.005804195804195804,
+      "grad_norm": 0.8167915940284729,
+      "learning_rate": 8.706293706293705e-05,
+      "loss": 7.0738,
+      "step": 83
+    },
+    {
+      "epoch": 0.005874125874125874,
+      "grad_norm": 0.6965641975402832,
+      "learning_rate": 8.81118881118881e-05,
+      "loss": 7.1085,
+      "step": 84
+    },
+    {
+      "epoch": 0.005944055944055944,
+      "grad_norm": 0.9094026684761047,
+      "learning_rate": 8.916083916083914e-05,
+      "loss": 7.0906,
+      "step": 85
+    },
+    {
+      "epoch": 0.006013986013986014,
+      "grad_norm": 0.7994691133499146,
+      "learning_rate": 9.02097902097902e-05,
+      "loss": 7.0782,
+      "step": 86
+    },
+    {
+      "epoch": 0.006083916083916084,
+      "grad_norm": 1.007351279258728,
+      "learning_rate": 9.125874125874126e-05,
+      "loss": 7.0095,
+      "step": 87
+    },
+    {
+      "epoch": 0.006153846153846154,
+      "grad_norm": 0.8736310601234436,
+      "learning_rate": 9.23076923076923e-05,
+      "loss": 6.9629,
+      "step": 88
+    },
+    {
+      "epoch": 0.0062237762237762236,
+      "grad_norm": 0.6392287015914917,
+      "learning_rate": 9.335664335664336e-05,
+      "loss": 6.9908,
+      "step": 89
+    },
+    {
+      "epoch": 0.006293706293706294,
+      "grad_norm": 1.091141939163208,
+      "learning_rate": 9.44055944055944e-05,
+      "loss": 6.9523,
+      "step": 90
+    },
+    {
+      "epoch": 0.006363636363636364,
+      "grad_norm": 0.7908533215522766,
+      "learning_rate": 9.545454545454545e-05,
+      "loss": 6.9238,
+      "step": 91
+    },
+    {
+      "epoch": 0.0064335664335664336,
+      "grad_norm": 0.6706556081771851,
+      "learning_rate": 9.650349650349649e-05,
+      "loss": 6.9672,
+      "step": 92
+    },
+    {
+      "epoch": 0.006503496503496503,
+      "grad_norm": 0.9139024019241333,
+      "learning_rate": 9.755244755244754e-05,
+      "loss": 6.9997,
+      "step": 93
+    },
+    {
+      "epoch": 0.006573426573426574,
+      "grad_norm": 0.7192760109901428,
+      "learning_rate": 9.860139860139858e-05,
+      "loss": 6.936,
+      "step": 94
+    },
+    {
+      "epoch": 0.006643356643356644,
+      "grad_norm": 0.7734940052032471,
+      "learning_rate": 9.965034965034964e-05,
+      "loss": 7.0156,
+      "step": 95
+    },
+    {
+      "epoch": 0.006713286713286713,
+      "grad_norm": 0.8069332838058472,
+      "learning_rate": 0.0001006993006993007,
+      "loss": 6.8958,
+      "step": 96
+    },
+    {
+      "epoch": 0.006783216783216783,
+      "grad_norm": 1.0324732065200806,
+      "learning_rate": 0.00010174825174825174,
+      "loss": 6.8863,
+      "step": 97
+    },
+    {
+      "epoch": 0.006853146853146854,
+      "grad_norm": 0.6798093318939209,
+      "learning_rate": 0.00010279720279720279,
+      "loss": 6.8534,
+      "step": 98
+    },
+    {
+      "epoch": 0.006923076923076923,
+      "grad_norm": 0.7703111171722412,
+      "learning_rate": 0.00010384615384615383,
+      "loss": 6.7499,
+      "step": 99
+    },
+    {
+      "epoch": 0.006993006993006993,
+      "grad_norm": 0.941916823387146,
+      "learning_rate": 0.00010489510489510488,
+      "loss": 6.8574,
+      "step": 100
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 14300,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 9223372036854775807,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.7825750933504e+16,
+  "train_batch_size": 64,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-100/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d8b5ab7c96eddd31884a55cca6ae1cb4177ea7db30f3aa145cab6706fc99c36
+size 7288

checkpoint-100/zero_to_fp32.py

@@ -0,0 +1,760 @@
+#!/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 . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import gc
+import json
+import numpy as np
+from tqdm import tqdm
+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, weights_only=False)
+
+        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 tqdm(files, desc='Loading checkpoint shards'):
+        state_dict = torch.load(f, map_location=device, mmap=True, weights_only=False)
+        # 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}")
+
+    fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] 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")
+
+
+class GatheredTensor:
+    """
+    A pseudo tensor that collects partitioned weights.
+    It is more memory efficient when there are multiple groups.
+    """
+
+    def __init__(self, flat_groups, flat_groups_offset, offset, partitioned_numel, shape):
+        self.flat_groups = flat_groups
+        self.flat_groups_offset = flat_groups_offset
+        self.offset = offset
+        self.partitioned_numel = partitioned_numel
+        self.shape = shape
+        self.dtype = self.flat_groups[0][0].dtype
+
+    def contiguous(self):
+        """
+        Merge partitioned weights from flat_groups into a single tensor.
+        """
+        end_idx = self.offset + self.partitioned_numel
+        world_size = len(self.flat_groups)
+        pad_flat_param_chunks = []
+
+        for rank_i in range(world_size):
+            # for each rank, we need to collect weights from related group/groups
+            flat_groups_at_rank_i = self.flat_groups[rank_i]
+            start_group_id = None
+            end_group_id = None
+            for group_id in range(len(self.flat_groups_offset)):
+                if self.flat_groups_offset[group_id] <= self.offset < self.flat_groups_offset[group_id + 1]:
+                    start_group_id = group_id
+                if self.flat_groups_offset[group_id] < end_idx <= self.flat_groups_offset[group_id + 1]:
+                    end_group_id = group_id
+                    break
+            # collect weights from related group/groups
+            for group_id in range(start_group_id, end_group_id + 1):
+                flat_tensor = flat_groups_at_rank_i[group_id]
+                start_offset = self.offset - self.flat_groups_offset[group_id]
+                end_offset = min(end_idx, self.flat_groups_offset[group_id + 1]) - self.flat_groups_offset[group_id]
+                pad_flat_param_chunks.append(flat_tensor[start_offset:end_offset])
+
+        # collect weights from all ranks
+        pad_flat_param = torch.cat(pad_flat_param_chunks, dim=0)
+        param = pad_flat_param[:self.shape.numel()].view(self.shape).contiguous()
+        return param
+
+
+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 = sum([flat_group.numel() for flat_group in fp32_flat_groups[0]]) * 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
+    flat_groups_offset = [0] + list(np.cumsum([flat_tensor.numel() for flat_tensor in fp32_flat_groups[0]]))
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering sharded weights'):
+        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}"
+            )
+
+        # memory efficient tensor
+        tensor = GatheredTensor(fp32_flat_groups, flat_groups_offset, offset, partitioned_numel, shape)
+        state_dict[name] = tensor
+        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 to_torch_tensor(state_dict, return_empty_tensor=False):
+    """
+    Convert state_dict of GatheredTensor to torch tensor
+    """
+    torch_state_dict = {}
+    converted_tensors = {}
+    for name, tensor in state_dict.items():
+        tensor_id = id(tensor)
+        if tensor_id in converted_tensors:  # shared tensors
+            shared_tensor = torch_state_dict[converted_tensors[tensor_id]]
+            torch_state_dict[name] = shared_tensor
+        else:
+            converted_tensors[tensor_id] = name
+            if return_empty_tensor:
+                torch_state_dict[name] = torch.empty(tensor.shape, dtype=tensor.dtype)
+            else:
+                torch_state_dict[name] = tensor.contiguous()
+    return torch_state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                             tag=None,
+                                             exclude_frozen_parameters=False,
+                                             lazy_mode=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
+        - ``lazy_mode``: get state_dict in lazy mode. It returns a dict of pesduo tensor instead of torch tensor, which is more memory efficient.
+          Convert the pesduo tensor to torch tensor by ``.contiguous()``
+
+    Returns:
+        - pytorch ``state_dict``
+
+    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.
+
+    Note: the above usage may not work if your application doesn't have sufficient free CPU memory.
+    You may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint. Or you can load state_dict in lazy mode ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, lazy_mode=True) # not on cpu
+        for name, lazy_tensor in state_dict.item():
+            tensor = lazy_tensor.contiguous()  # to cpu
+            print(name, tensor)
+            # del tensor to release memory if it no longer in use
+    """
+    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")
+
+    state_dict = _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+    if lazy_mode:
+        return state_dict
+    else:
+        return to_torch_tensor(state_dict)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               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_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``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
+    """
+
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                                          tag,
+                                                          exclude_frozen_parameters,
+                                                          lazy_mode=True)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        # an memory-efficient approach for sharding
+        empty_state_dict = to_torch_tensor(state_dict, return_empty_tensor=True)
+        state_dict_split = split_torch_state_dict_into_shards(empty_state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model by shard
+    os.makedirs(output_dir, exist_ok=True)
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard_state_dict = {tensor_name: state_dict[tensor_name] for tensor_name in tensors}
+        shard_state_dict = to_torch_tensor(shard_state_dict)
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard_state_dict, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard_state_dict, output_path)
+        # release the memory of current shard
+        for tensor_name in list(shard_state_dict.keys()):
+            del state_dict[tensor_name]
+            del shard_state_dict[tensor_name]
+        del shard_state_dict
+        gc.collect()
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+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_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    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_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)