upload weights

config.json

@@ -0,0 +1,41 @@
+{
+  "_name_or_path": "EleutherAI/gpt-j-6B",
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPTJForCausalLM"
+  ],
+  "attn_pdrop": 0.0,
+  "bos_token_id": 50256,
+  "embd_pdrop": 0.0,
+  "eos_token_id": 50256,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gptj",
+  "n_embd": 4096,
+  "n_head": 16,
+  "n_inner": null,
+  "n_layer": 28,
+  "n_positions": 2048,
+  "resid_pdrop": 0.0,
+  "rotary": true,
+  "rotary_dim": 64,
+  "scale_attn_weights": true,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "task_specific_params": {
+    "text-generation": {
+      "do_sample": true,
+      "max_length": 50,
+      "temperature": 1.0
+    }
+  },
+  "tie_word_embeddings": false,
+  "tokenizer_class": "GPT2Tokenizer",
+  "torch_dtype": "float16",
+  "transformers_version": "4.28.1",
+  "use_cache": false,
+  "vocab_size": 50257
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 50256,
+  "eos_token_id": 50256,
+  "transformers_version": "4.28.1",
+  "use_cache": false
+}

latest

@@ -0,0 +1 @@
+global_step680

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b81fe246fe829ef1903ad6d268793f05ac9bf3fd03cef393ca5a0eb22cc89e52
+size 12216972905

rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0b47786a021fed7e4aacc8d852ba10a39c20411f6f2cf6d8ffc637f14955a74
+size 14583

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9ebaeb0d5485091d948a19f87ccdc8f7dc709b1c2f9d7386c9ee19f298e132d
+size 14583

rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:766323763276f8db8b322269cb86e0e257714803820af7a0f4297c2b0b635806
+size 14583

rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbddaf920d7cbb89da457891cc70e8694977c94948058fb63294c63f16f5b5ba
+size 14583

special_tokens_map.json

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

tokenizer_config.json

@@ -0,0 +1,33 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "bos_token": {
+    "__type": "AddedToken",
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "clean_up_tokenization_spaces": true,
+  "eos_token": {
+    "__type": "AddedToken",
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "errors": "replace",
+  "model_max_length": 2048,
+  "pad_token": null,
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": {
+    "__type": "AddedToken",
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  }
+}

trainer_state.json

@@ -0,0 +1,294 @@
+{
+  "best_metric": 0.057281494140625,
+  "best_model_checkpoint": "/opt/ml/checkpoints/checkpoint-680",
+  "epoch": 0.26143790849673204,
+  "global_step": 680,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "eval_loss": 0.07366943359375,
+      "eval_runtime": 1988.0122,
+      "eval_samples_per_second": 5.232,
+      "eval_steps_per_second": 0.164,
+      "step": 20
+    },
+    {
+      "epoch": 0.02,
+      "eval_loss": 0.06402587890625,
+      "eval_runtime": 1988.5161,
+      "eval_samples_per_second": 5.231,
+      "eval_steps_per_second": 0.164,
+      "step": 40
+    },
+    {
+      "epoch": 0.02,
+      "eval_loss": 0.060882568359375,
+      "eval_runtime": 1988.0139,
+      "eval_samples_per_second": 5.232,
+      "eval_steps_per_second": 0.164,
+      "step": 60
+    },
+    {
+      "epoch": 0.03,
+      "eval_loss": 0.05999755859375,
+      "eval_runtime": 1986.9046,
+      "eval_samples_per_second": 5.235,
+      "eval_steps_per_second": 0.164,
+      "step": 80
+    },
+    {
+      "epoch": 0.04,
+      "eval_loss": 0.059539794921875,
+      "eval_runtime": 1986.5359,
+      "eval_samples_per_second": 5.236,
+      "eval_steps_per_second": 0.164,
+      "step": 100
+    },
+    {
+      "epoch": 0.05,
+      "eval_loss": 0.05914306640625,
+      "eval_runtime": 1986.2404,
+      "eval_samples_per_second": 5.237,
+      "eval_steps_per_second": 0.164,
+      "step": 120
+    },
+    {
+      "epoch": 0.05,
+      "eval_loss": 0.0589599609375,
+      "eval_runtime": 1985.8515,
+      "eval_samples_per_second": 5.238,
+      "eval_steps_per_second": 0.164,
+      "step": 140
+    },
+    {
+      "epoch": 0.06,
+      "eval_loss": 0.058807373046875,
+      "eval_runtime": 1985.7582,
+      "eval_samples_per_second": 5.238,
+      "eval_steps_per_second": 0.164,
+      "step": 160
+    },
+    {
+      "epoch": 0.07,
+      "eval_loss": 0.05853271484375,
+      "eval_runtime": 1986.2944,
+      "eval_samples_per_second": 5.236,
+      "eval_steps_per_second": 0.164,
+      "step": 180
+    },
+    {
+      "epoch": 0.08,
+      "eval_loss": 0.05841064453125,
+      "eval_runtime": 1986.0105,
+      "eval_samples_per_second": 5.237,
+      "eval_steps_per_second": 0.164,
+      "step": 200
+    },
+    {
+      "epoch": 0.08,
+      "eval_loss": 0.058349609375,
+      "eval_runtime": 1986.7151,
+      "eval_samples_per_second": 5.235,
+      "eval_steps_per_second": 0.164,
+      "step": 220
+    },
+    {
+      "epoch": 0.09,
+      "eval_loss": 0.058197021484375,
+      "eval_runtime": 1985.7618,
+      "eval_samples_per_second": 5.238,
+      "eval_steps_per_second": 0.164,
+      "step": 240
+    },
+    {
+      "epoch": 0.1,
+      "eval_loss": 0.058135986328125,
+      "eval_runtime": 1985.9202,
+      "eval_samples_per_second": 5.237,
+      "eval_steps_per_second": 0.164,
+      "step": 260
+    },
+    {
+      "epoch": 0.11,
+      "eval_loss": 0.058135986328125,
+      "eval_runtime": 1986.6184,
+      "eval_samples_per_second": 5.236,
+      "eval_steps_per_second": 0.164,
+      "step": 280
+    },
+    {
+      "epoch": 0.12,
+      "eval_loss": 0.0579833984375,
+      "eval_runtime": 1986.6502,
+      "eval_samples_per_second": 5.235,
+      "eval_steps_per_second": 0.164,
+      "step": 300
+    },
+    {
+      "epoch": 0.12,
+      "eval_loss": 0.0579833984375,
+      "eval_runtime": 1985.5739,
+      "eval_samples_per_second": 5.238,
+      "eval_steps_per_second": 0.164,
+      "step": 320
+    },
+    {
+      "epoch": 0.13,
+      "eval_loss": 0.057861328125,
+      "eval_runtime": 1985.7642,
+      "eval_samples_per_second": 5.238,
+      "eval_steps_per_second": 0.164,
+      "step": 340
+    },
+    {
+      "epoch": 0.14,
+      "eval_loss": 0.0577392578125,
+      "eval_runtime": 1985.2011,
+      "eval_samples_per_second": 5.239,
+      "eval_steps_per_second": 0.164,
+      "step": 360
+    },
+    {
+      "epoch": 0.15,
+      "eval_loss": 0.057769775390625,
+      "eval_runtime": 1985.6821,
+      "eval_samples_per_second": 5.238,
+      "eval_steps_per_second": 0.164,
+      "step": 380
+    },
+    {
+      "epoch": 0.15,
+      "eval_loss": 0.057647705078125,
+      "eval_runtime": 1986.1381,
+      "eval_samples_per_second": 5.237,
+      "eval_steps_per_second": 0.164,
+      "step": 400
+    },
+    {
+      "epoch": 0.16,
+      "eval_loss": 0.05767822265625,
+      "eval_runtime": 1985.6719,
+      "eval_samples_per_second": 5.238,
+      "eval_steps_per_second": 0.164,
+      "step": 420
+    },
+    {
+      "epoch": 0.17,
+      "eval_loss": 0.057647705078125,
+      "eval_runtime": 1985.643,
+      "eval_samples_per_second": 5.238,
+      "eval_steps_per_second": 0.164,
+      "step": 440
+    },
+    {
+      "epoch": 0.18,
+      "eval_loss": 0.057525634765625,
+      "eval_runtime": 1985.6414,
+      "eval_samples_per_second": 5.238,
+      "eval_steps_per_second": 0.164,
+      "step": 460
+    },
+    {
+      "epoch": 0.18,
+      "eval_loss": 0.05755615234375,
+      "eval_runtime": 1985.6316,
+      "eval_samples_per_second": 5.238,
+      "eval_steps_per_second": 0.164,
+      "step": 480
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5e-06,
+      "loss": 0.2414,
+      "step": 500
+    },
+    {
+      "epoch": 0.19,
+      "eval_loss": 0.057525634765625,
+      "eval_runtime": 1986.8786,
+      "eval_samples_per_second": 5.235,
+      "eval_steps_per_second": 0.164,
+      "step": 500
+    },
+    {
+      "epoch": 0.2,
+      "eval_loss": 0.0574951171875,
+      "eval_runtime": 1987.3346,
+      "eval_samples_per_second": 5.234,
+      "eval_steps_per_second": 0.164,
+      "step": 520
+    },
+    {
+      "epoch": 0.21,
+      "eval_loss": 0.057464599609375,
+      "eval_runtime": 1986.2938,
+      "eval_samples_per_second": 5.236,
+      "eval_steps_per_second": 0.164,
+      "step": 540
+    },
+    {
+      "epoch": 0.22,
+      "eval_loss": 0.0574951171875,
+      "eval_runtime": 1986.0284,
+      "eval_samples_per_second": 5.237,
+      "eval_steps_per_second": 0.164,
+      "step": 560
+    },
+    {
+      "epoch": 0.22,
+      "eval_loss": 0.057403564453125,
+      "eval_runtime": 1987.0081,
+      "eval_samples_per_second": 5.235,
+      "eval_steps_per_second": 0.164,
+      "step": 580
+    },
+    {
+      "epoch": 0.23,
+      "eval_loss": 0.057403564453125,
+      "eval_runtime": 1987.2441,
+      "eval_samples_per_second": 5.234,
+      "eval_steps_per_second": 0.164,
+      "step": 600
+    },
+    {
+      "epoch": 0.24,
+      "eval_loss": 0.057403564453125,
+      "eval_runtime": 1988.9229,
+      "eval_samples_per_second": 5.229,
+      "eval_steps_per_second": 0.164,
+      "step": 620
+    },
+    {
+      "epoch": 0.25,
+      "eval_loss": 0.057403564453125,
+      "eval_runtime": 1988.9574,
+      "eval_samples_per_second": 5.229,
+      "eval_steps_per_second": 0.164,
+      "step": 640
+    },
+    {
+      "epoch": 0.25,
+      "eval_loss": 0.057342529296875,
+      "eval_runtime": 1989.2299,
+      "eval_samples_per_second": 5.229,
+      "eval_steps_per_second": 0.164,
+      "step": 660
+    },
+    {
+      "epoch": 0.26,
+      "eval_loss": 0.057281494140625,
+      "eval_runtime": 1989.7704,
+      "eval_samples_per_second": 5.227,
+      "eval_steps_per_second": 0.164,
+      "step": 680
+    }
+  ],
+  "max_steps": 31212,
+  "num_train_epochs": 12,
+  "total_flos": 108266998726656.0,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a9191e8a5e0bb4e333a1882c36df2f9a5b47697d3ecd47a8f9a1ffca648d116
+size 4923

zero_to_fp32.py

@@ -0,0 +1,483 @@
+#!/usr/bin/env python
+'''Copyright The Microsoft DeepSpeed Team'''
+
+# This script extracts fp32 consolidated weights from a zero 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
+
+# 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)
+
+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_optim_files(checkpoint_dir):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    optim_files = sorted(glob.glob(os.path.join(checkpoint_dir,
+                                                "*_optim_states.pt")),
+                         key=natural_keys)
+
+    if len(optim_files) == 0:
+        raise FileNotFoundError(
+            f"can't find '*_optim_states.pt' files in directory '{checkpoint_dir}'")
+
+    return optim_files
+
+
+def parse_model_state(file):
+    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]
+
+    ds_version = state_dict.get(DS_VERSION, None)
+
+    return buffers, param_shapes, ds_version
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dicts.append(torch.load(f, map_location=device))
+
+    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):
+    """
+    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_file = get_model_state_file(ds_checkpoint_dir, zero_stage)
+    buffers, param_shapes, ds_version = parse_model_state(model_file)
+    print(f'Parsing checkpoint created by deepspeed=={ds_version}')
+
+    if zero_stage == 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size,
+                                                          param_shapes,
+                                                          fp32_flat_groups,
+                                                          buffers)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size,
+                                                          param_shapes,
+                                                          fp32_flat_groups,
+                                                          buffers)
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size,
+                                               param_shapes,
+                                               fp32_flat_groups,
+                                               buffers):
+
+    # 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.")
+
+    state_dict = OrderedDict()
+
+    # buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    # 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()
+            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"
+    )
+
+    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 _get_fp32_state_dict_from_zero3_checkpoint(world_size,
+                                               param_shapes,
+                                               fp32_flat_groups,
+                                               buffers):
+
+    # 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
+
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # 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
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    state_dict = OrderedDict()
+
+    # buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    # 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"{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 fp32 state dict with {total_params} params {total_numel} elements"
+    )
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
+    """
+    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``
+
+    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)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
+    """
+    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``
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+    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("-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)