Upload folder using huggingface_hub

README.md

@@ -0,0 +1,202 @@
+---
+base_model: meta-llama/Llama-3.1-8B
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.13.2

adapter_config.json

@@ -0,0 +1,36 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "meta-llama/Llama-3.1-8B",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 64,
+  "lora_dropout": 0.1,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": [
+    "classifier",
+    "score"
+  ],
+  "peft_type": "LORA",
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "down_proj",
+    "q_proj",
+    "up_proj",
+    "v_proj",
+    "o_proj"
+  ],
+  "task_type": "SEQ_CLS",
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f52fd3fd3df18dd00c06b077f84d09a1b77d968cf47b6416083b8ce63023f6a5
+size 157428238

latest

@@ -0,0 +1 @@
+global_step1000

rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cae35d58c6f0a96413f50b5f6df11c4603e5055bfae9395f74b37d830c668519
+size 15024

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbcbe4dffbaae5e4c54a83d5b216d144936bc304cd200ea4d9bef91915c2411f
+size 14960

rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f2fefd72a6fe9c258605b37db501943def18c128791b5b17ff7b07509e1dfb8
+size 15024

rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4af83f1b60de97151e3a69400730cfdb7045fdca693d7cbbc3c825d5b5c882ed
+size 14960

trainer_state.json

@@ -0,0 +1,733 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.6387225548902196,
+  "eval_steps": 500,
+  "global_step": 1000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.006387225548902195,
+      "grad_norm": 3.5834698688009654,
+      "learning_rate": 4.007141788485205e-05,
+      "loss": 0.8162,
+      "step": 10
+    },
+    {
+      "epoch": 0.01277445109780439,
+      "grad_norm": 3.7982361966359415,
+      "learning_rate": 5.213411663697864e-05,
+      "loss": 0.4857,
+      "step": 20
+    },
+    {
+      "epoch": 0.019161676646706587,
+      "grad_norm": 3.3097747900168066,
+      "learning_rate": 5.919034306446858e-05,
+      "loss": 0.1919,
+      "step": 30
+    },
+    {
+      "epoch": 0.02554890219560878,
+      "grad_norm": 1.6554650990460391,
+      "learning_rate": 6.419681538910523e-05,
+      "loss": 0.1177,
+      "step": 40
+    },
+    {
+      "epoch": 0.031936127744510975,
+      "grad_norm": 1.0526158883209893,
+      "learning_rate": 6.80801370175775e-05,
+      "loss": 0.0997,
+      "step": 50
+    },
+    {
+      "epoch": 0.03832335329341317,
+      "grad_norm": 1.322844958779453,
+      "learning_rate": 7.125304181659517e-05,
+      "loss": 0.0987,
+      "step": 60
+    },
+    {
+      "epoch": 0.04471057884231537,
+      "grad_norm": 0.906098163414632,
+      "learning_rate": 7.393569459993276e-05,
+      "loss": 0.0902,
+      "step": 70
+    },
+    {
+      "epoch": 0.05109780439121756,
+      "grad_norm": 0.9742791206427208,
+      "learning_rate": 7.625951414123182e-05,
+      "loss": 0.08,
+      "step": 80
+    },
+    {
+      "epoch": 0.05748502994011976,
+      "grad_norm": 0.7369633538190276,
+      "learning_rate": 7.83092682440851e-05,
+      "loss": 0.0598,
+      "step": 90
+    },
+    {
+      "epoch": 0.06387225548902195,
+      "grad_norm": 0.6187428023320575,
+      "learning_rate": 8.01428357697041e-05,
+      "loss": 0.0836,
+      "step": 100
+    },
+    {
+      "epoch": 0.07025948103792415,
+      "grad_norm": 0.7404818716311581,
+      "learning_rate": 8.180149935405545e-05,
+      "loss": 0.0758,
+      "step": 110
+    },
+    {
+      "epoch": 0.07664670658682635,
+      "grad_norm": 0.4603749462575225,
+      "learning_rate": 8.331574056872175e-05,
+      "loss": 0.0698,
+      "step": 120
+    },
+    {
+      "epoch": 0.08303393213572854,
+      "grad_norm": 0.5780703019259222,
+      "learning_rate": 8.470870745519228e-05,
+      "loss": 0.0771,
+      "step": 130
+    },
+    {
+      "epoch": 0.08942115768463074,
+      "grad_norm": 0.7869338127501012,
+      "learning_rate": 8.599839335205934e-05,
+      "loss": 0.0741,
+      "step": 140
+    },
+    {
+      "epoch": 0.09580838323353294,
+      "grad_norm": 0.5085534390538137,
+      "learning_rate": 8.719906219719403e-05,
+      "loss": 0.0637,
+      "step": 150
+    },
+    {
+      "epoch": 0.10219560878243512,
+      "grad_norm": 0.6419093548122555,
+      "learning_rate": 8.83222128933584e-05,
+      "loss": 0.0683,
+      "step": 160
+    },
+    {
+      "epoch": 0.10858283433133732,
+      "grad_norm": 0.4014377657568054,
+      "learning_rate": 8.937725079936634e-05,
+      "loss": 0.0732,
+      "step": 170
+    },
+    {
+      "epoch": 0.11497005988023952,
+      "grad_norm": 0.4686184226271436,
+      "learning_rate": 9.03719669962117e-05,
+      "loss": 0.0727,
+      "step": 180
+    },
+    {
+      "epoch": 0.12135728542914172,
+      "grad_norm": 0.3553126799550481,
+      "learning_rate": 9.13128878003922e-05,
+      "loss": 0.0642,
+      "step": 190
+    },
+    {
+      "epoch": 0.1277445109780439,
+      "grad_norm": 0.2308342524288462,
+      "learning_rate": 9.220553452183068e-05,
+      "loss": 0.0662,
+      "step": 200
+    },
+    {
+      "epoch": 0.1341317365269461,
+      "grad_norm": 0.4244978948516096,
+      "learning_rate": 9.305461977954928e-05,
+      "loss": 0.0653,
+      "step": 210
+    },
+    {
+      "epoch": 0.1405189620758483,
+      "grad_norm": 0.6491159925885343,
+      "learning_rate": 9.386419810618205e-05,
+      "loss": 0.0692,
+      "step": 220
+    },
+    {
+      "epoch": 0.1469061876247505,
+      "grad_norm": 0.6586709935585672,
+      "learning_rate": 9.46377830473483e-05,
+      "loss": 0.0603,
+      "step": 230
+    },
+    {
+      "epoch": 0.1532934131736527,
+      "grad_norm": 0.4334717142812121,
+      "learning_rate": 9.537843932084834e-05,
+      "loss": 0.075,
+      "step": 240
+    },
+    {
+      "epoch": 0.1596806387225549,
+      "grad_norm": 0.31119556502835766,
+      "learning_rate": 9.608885615030295e-05,
+      "loss": 0.0646,
+      "step": 250
+    },
+    {
+      "epoch": 0.1660678642714571,
+      "grad_norm": 0.3201297094259269,
+      "learning_rate": 9.677140620731887e-05,
+      "loss": 0.0687,
+      "step": 260
+    },
+    {
+      "epoch": 0.1724550898203593,
+      "grad_norm": 0.3607671741835552,
+      "learning_rate": 9.742819342370165e-05,
+      "loss": 0.0611,
+      "step": 270
+    },
+    {
+      "epoch": 0.17884231536926148,
+      "grad_norm": 0.2599965595224881,
+      "learning_rate": 9.806109210418593e-05,
+      "loss": 0.0682,
+      "step": 280
+    },
+    {
+      "epoch": 0.18522954091816368,
+      "grad_norm": 0.32241547083298766,
+      "learning_rate": 9.86717791726321e-05,
+      "loss": 0.0612,
+      "step": 290
+    },
+    {
+      "epoch": 0.19161676646706588,
+      "grad_norm": 0.48624224094653656,
+      "learning_rate": 9.926176094932063e-05,
+      "loss": 0.0569,
+      "step": 300
+    },
+    {
+      "epoch": 0.19800399201596808,
+      "grad_norm": 0.3035297139442493,
+      "learning_rate": 9.983239553594597e-05,
+      "loss": 0.0581,
+      "step": 310
+    },
+    {
+      "epoch": 0.20439121756487025,
+      "grad_norm": 0.2209458722911404,
+      "learning_rate": 9.97870074547391e-05,
+      "loss": 0.0579,
+      "step": 320
+    },
+    {
+      "epoch": 0.21077844311377245,
+      "grad_norm": 0.32849509401911653,
+      "learning_rate": 9.943201987930422e-05,
+      "loss": 0.0594,
+      "step": 330
+    },
+    {
+      "epoch": 0.21716566866267464,
+      "grad_norm": 0.3121594199523004,
+      "learning_rate": 9.907703230386936e-05,
+      "loss": 0.0689,
+      "step": 340
+    },
+    {
+      "epoch": 0.22355289421157684,
+      "grad_norm": 0.37282218523682487,
+      "learning_rate": 9.87220447284345e-05,
+      "loss": 0.0591,
+      "step": 350
+    },
+    {
+      "epoch": 0.22994011976047904,
+      "grad_norm": 0.21932185059016668,
+      "learning_rate": 9.836705715299965e-05,
+      "loss": 0.0561,
+      "step": 360
+    },
+    {
+      "epoch": 0.23632734530938124,
+      "grad_norm": 0.40431591808333317,
+      "learning_rate": 9.801206957756479e-05,
+      "loss": 0.0547,
+      "step": 370
+    },
+    {
+      "epoch": 0.24271457085828343,
+      "grad_norm": 0.3545382413148979,
+      "learning_rate": 9.765708200212993e-05,
+      "loss": 0.068,
+      "step": 380
+    },
+    {
+      "epoch": 0.24910179640718563,
+      "grad_norm": 0.3708116952070532,
+      "learning_rate": 9.730209442669507e-05,
+      "loss": 0.0644,
+      "step": 390
+    },
+    {
+      "epoch": 0.2554890219560878,
+      "grad_norm": 0.2933151282788518,
+      "learning_rate": 9.69471068512602e-05,
+      "loss": 0.0554,
+      "step": 400
+    },
+    {
+      "epoch": 0.26187624750499,
+      "grad_norm": 0.2701122564574567,
+      "learning_rate": 9.659211927582535e-05,
+      "loss": 0.0669,
+      "step": 410
+    },
+    {
+      "epoch": 0.2682634730538922,
+      "grad_norm": 0.43233913171276295,
+      "learning_rate": 9.62371317003905e-05,
+      "loss": 0.0696,
+      "step": 420
+    },
+    {
+      "epoch": 0.2746506986027944,
+      "grad_norm": 0.29840407730897334,
+      "learning_rate": 9.588214412495564e-05,
+      "loss": 0.0583,
+      "step": 430
+    },
+    {
+      "epoch": 0.2810379241516966,
+      "grad_norm": 0.347519310387429,
+      "learning_rate": 9.552715654952076e-05,
+      "loss": 0.0675,
+      "step": 440
+    },
+    {
+      "epoch": 0.2874251497005988,
+      "grad_norm": 0.23159546291731048,
+      "learning_rate": 9.51721689740859e-05,
+      "loss": 0.0595,
+      "step": 450
+    },
+    {
+      "epoch": 0.293812375249501,
+      "grad_norm": 0.19539165396565417,
+      "learning_rate": 9.481718139865106e-05,
+      "loss": 0.0593,
+      "step": 460
+    },
+    {
+      "epoch": 0.3001996007984032,
+      "grad_norm": 0.3711679298627976,
+      "learning_rate": 9.446219382321619e-05,
+      "loss": 0.0576,
+      "step": 470
+    },
+    {
+      "epoch": 0.3065868263473054,
+      "grad_norm": 0.21594608036662566,
+      "learning_rate": 9.410720624778133e-05,
+      "loss": 0.0612,
+      "step": 480
+    },
+    {
+      "epoch": 0.3129740518962076,
+      "grad_norm": 0.38847976792533884,
+      "learning_rate": 9.375221867234647e-05,
+      "loss": 0.0558,
+      "step": 490
+    },
+    {
+      "epoch": 0.3193612774451098,
+      "grad_norm": 0.48499489939381346,
+      "learning_rate": 9.339723109691161e-05,
+      "loss": 0.0598,
+      "step": 500
+    },
+    {
+      "epoch": 0.325748502994012,
+      "grad_norm": 0.26871455217005846,
+      "learning_rate": 9.304224352147675e-05,
+      "loss": 0.0599,
+      "step": 510
+    },
+    {
+      "epoch": 0.3321357285429142,
+      "grad_norm": 0.3006779859330303,
+      "learning_rate": 9.26872559460419e-05,
+      "loss": 0.0533,
+      "step": 520
+    },
+    {
+      "epoch": 0.3385229540918164,
+      "grad_norm": 0.252910708387006,
+      "learning_rate": 9.233226837060704e-05,
+      "loss": 0.0609,
+      "step": 530
+    },
+    {
+      "epoch": 0.3449101796407186,
+      "grad_norm": 0.28615902583412134,
+      "learning_rate": 9.197728079517217e-05,
+      "loss": 0.0632,
+      "step": 540
+    },
+    {
+      "epoch": 0.35129740518962077,
+      "grad_norm": 0.3309472244637015,
+      "learning_rate": 9.162229321973731e-05,
+      "loss": 0.0657,
+      "step": 550
+    },
+    {
+      "epoch": 0.35768463073852297,
+      "grad_norm": 0.3181277078140241,
+      "learning_rate": 9.126730564430246e-05,
+      "loss": 0.0589,
+      "step": 560
+    },
+    {
+      "epoch": 0.36407185628742517,
+      "grad_norm": 0.5196112858300266,
+      "learning_rate": 9.09123180688676e-05,
+      "loss": 0.0645,
+      "step": 570
+    },
+    {
+      "epoch": 0.37045908183632736,
+      "grad_norm": 0.4316495610405202,
+      "learning_rate": 9.055733049343273e-05,
+      "loss": 0.0659,
+      "step": 580
+    },
+    {
+      "epoch": 0.37684630738522956,
+      "grad_norm": 0.25853510285529524,
+      "learning_rate": 9.020234291799787e-05,
+      "loss": 0.0582,
+      "step": 590
+    },
+    {
+      "epoch": 0.38323353293413176,
+      "grad_norm": 0.28576780733983087,
+      "learning_rate": 8.984735534256301e-05,
+      "loss": 0.0531,
+      "step": 600
+    },
+    {
+      "epoch": 0.38962075848303396,
+      "grad_norm": 0.30628185946436504,
+      "learning_rate": 8.949236776712816e-05,
+      "loss": 0.0577,
+      "step": 610
+    },
+    {
+      "epoch": 0.39600798403193616,
+      "grad_norm": 0.22146965371135377,
+      "learning_rate": 8.91373801916933e-05,
+      "loss": 0.0571,
+      "step": 620
+    },
+    {
+      "epoch": 0.4023952095808383,
+      "grad_norm": 0.2720642528700885,
+      "learning_rate": 8.878239261625844e-05,
+      "loss": 0.059,
+      "step": 630
+    },
+    {
+      "epoch": 0.4087824351297405,
+      "grad_norm": 0.35619962554831525,
+      "learning_rate": 8.842740504082358e-05,
+      "loss": 0.0614,
+      "step": 640
+    },
+    {
+      "epoch": 0.4151696606786427,
+      "grad_norm": 0.28668258535201413,
+      "learning_rate": 8.807241746538871e-05,
+      "loss": 0.0699,
+      "step": 650
+    },
+    {
+      "epoch": 0.4215568862275449,
+      "grad_norm": 0.35772133791366534,
+      "learning_rate": 8.771742988995385e-05,
+      "loss": 0.0599,
+      "step": 660
+    },
+    {
+      "epoch": 0.4279441117764471,
+      "grad_norm": 0.3547047177760543,
+      "learning_rate": 8.7362442314519e-05,
+      "loss": 0.0563,
+      "step": 670
+    },
+    {
+      "epoch": 0.4343313373253493,
+      "grad_norm": 0.24898549550006935,
+      "learning_rate": 8.700745473908413e-05,
+      "loss": 0.0553,
+      "step": 680
+    },
+    {
+      "epoch": 0.4407185628742515,
+      "grad_norm": 0.34439451125458864,
+      "learning_rate": 8.665246716364927e-05,
+      "loss": 0.0525,
+      "step": 690
+    },
+    {
+      "epoch": 0.4471057884231537,
+      "grad_norm": 0.27788279819961054,
+      "learning_rate": 8.629747958821441e-05,
+      "loss": 0.0559,
+      "step": 700
+    },
+    {
+      "epoch": 0.4534930139720559,
+      "grad_norm": 0.358392882540888,
+      "learning_rate": 8.594249201277956e-05,
+      "loss": 0.0657,
+      "step": 710
+    },
+    {
+      "epoch": 0.4598802395209581,
+      "grad_norm": 0.3111624222426367,
+      "learning_rate": 8.55875044373447e-05,
+      "loss": 0.0472,
+      "step": 720
+    },
+    {
+      "epoch": 0.4662674650698603,
+      "grad_norm": 0.26806713373994445,
+      "learning_rate": 8.523251686190984e-05,
+      "loss": 0.0696,
+      "step": 730
+    },
+    {
+      "epoch": 0.4726546906187625,
+      "grad_norm": 0.2748488176683564,
+      "learning_rate": 8.487752928647498e-05,
+      "loss": 0.063,
+      "step": 740
+    },
+    {
+      "epoch": 0.47904191616766467,
+      "grad_norm": 0.18592617627322836,
+      "learning_rate": 8.452254171104012e-05,
+      "loss": 0.0609,
+      "step": 750
+    },
+    {
+      "epoch": 0.48542914171656687,
+      "grad_norm": 0.38695621265068386,
+      "learning_rate": 8.416755413560525e-05,
+      "loss": 0.062,
+      "step": 760
+    },
+    {
+      "epoch": 0.49181636726546907,
+      "grad_norm": 0.3158810203428474,
+      "learning_rate": 8.38125665601704e-05,
+      "loss": 0.0592,
+      "step": 770
+    },
+    {
+      "epoch": 0.49820359281437127,
+      "grad_norm": 0.3362065389804566,
+      "learning_rate": 8.345757898473555e-05,
+      "loss": 0.0613,
+      "step": 780
+    },
+    {
+      "epoch": 0.5045908183632735,
+      "grad_norm": 0.25748557706326924,
+      "learning_rate": 8.310259140930067e-05,
+      "loss": 0.0529,
+      "step": 790
+    },
+    {
+      "epoch": 0.5109780439121756,
+      "grad_norm": 0.29455626027596066,
+      "learning_rate": 8.274760383386582e-05,
+      "loss": 0.065,
+      "step": 800
+    },
+    {
+      "epoch": 0.5173652694610779,
+      "grad_norm": 0.2564019757489936,
+      "learning_rate": 8.239261625843096e-05,
+      "loss": 0.0579,
+      "step": 810
+    },
+    {
+      "epoch": 0.52375249500998,
+      "grad_norm": 0.228508320126866,
+      "learning_rate": 8.20376286829961e-05,
+      "loss": 0.0618,
+      "step": 820
+    },
+    {
+      "epoch": 0.5301397205588823,
+      "grad_norm": 0.17395934565113735,
+      "learning_rate": 8.168264110756124e-05,
+      "loss": 0.0601,
+      "step": 830
+    },
+    {
+      "epoch": 0.5365269461077844,
+      "grad_norm": 0.1845587781224566,
+      "learning_rate": 8.132765353212638e-05,
+      "loss": 0.0646,
+      "step": 840
+    },
+    {
+      "epoch": 0.5429141716566867,
+      "grad_norm": 0.219121974587572,
+      "learning_rate": 8.097266595669152e-05,
+      "loss": 0.0602,
+      "step": 850
+    },
+    {
+      "epoch": 0.5493013972055888,
+      "grad_norm": 0.3134533264569832,
+      "learning_rate": 8.061767838125665e-05,
+      "loss": 0.0593,
+      "step": 860
+    },
+    {
+      "epoch": 0.555688622754491,
+      "grad_norm": 0.36630204830289737,
+      "learning_rate": 8.02626908058218e-05,
+      "loss": 0.0606,
+      "step": 870
+    },
+    {
+      "epoch": 0.5620758483033932,
+      "grad_norm": 0.2390442131459619,
+      "learning_rate": 7.990770323038695e-05,
+      "loss": 0.0572,
+      "step": 880
+    },
+    {
+      "epoch": 0.5684630738522954,
+      "grad_norm": 0.2906363057717337,
+      "learning_rate": 7.955271565495209e-05,
+      "loss": 0.0687,
+      "step": 890
+    },
+    {
+      "epoch": 0.5748502994011976,
+      "grad_norm": 0.2853917682780398,
+      "learning_rate": 7.919772807951722e-05,
+      "loss": 0.0643,
+      "step": 900
+    },
+    {
+      "epoch": 0.5812375249500998,
+      "grad_norm": 0.3544408088733472,
+      "learning_rate": 7.884274050408236e-05,
+      "loss": 0.0574,
+      "step": 910
+    },
+    {
+      "epoch": 0.587624750499002,
+      "grad_norm": 0.24181631469575632,
+      "learning_rate": 7.848775292864751e-05,
+      "loss": 0.0697,
+      "step": 920
+    },
+    {
+      "epoch": 0.5940119760479042,
+      "grad_norm": 0.28668164498739,
+      "learning_rate": 7.813276535321264e-05,
+      "loss": 0.0641,
+      "step": 930
+    },
+    {
+      "epoch": 0.6003992015968064,
+      "grad_norm": 0.29519716447463473,
+      "learning_rate": 7.777777777777778e-05,
+      "loss": 0.0547,
+      "step": 940
+    },
+    {
+      "epoch": 0.6067864271457086,
+      "grad_norm": 0.2797455777496235,
+      "learning_rate": 7.742279020234292e-05,
+      "loss": 0.0622,
+      "step": 950
+    },
+    {
+      "epoch": 0.6131736526946108,
+      "grad_norm": 0.20698136186718125,
+      "learning_rate": 7.706780262690806e-05,
+      "loss": 0.0538,
+      "step": 960
+    },
+    {
+      "epoch": 0.619560878243513,
+      "grad_norm": 0.3029284687538603,
+      "learning_rate": 7.671281505147319e-05,
+      "loss": 0.0553,
+      "step": 970
+    },
+    {
+      "epoch": 0.6259481037924152,
+      "grad_norm": 0.3308461091763622,
+      "learning_rate": 7.635782747603835e-05,
+      "loss": 0.0592,
+      "step": 980
+    },
+    {
+      "epoch": 0.6323353293413174,
+      "grad_norm": 0.2652350297520287,
+      "learning_rate": 7.600283990060349e-05,
+      "loss": 0.0611,
+      "step": 990
+    },
+    {
+      "epoch": 0.6387225548902196,
+      "grad_norm": 0.20354784105739357,
+      "learning_rate": 7.564785232516862e-05,
+      "loss": 0.059,
+      "step": 1000
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 3130,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

zero_to_fp32.py

@@ -0,0 +1,674 @@
+#!/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 json
+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)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in 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}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_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)
+
+    # 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")
+        state_dict_split = split_torch_state_dict_into_shards(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
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # 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)