Add fine-tuned model files

README.md

@@ -0,0 +1,62 @@
+---
+library_name: peft
+license: other
+base_model: llava-hf/LLaVA-NeXT-Video-7B-hf
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+model-index:
+- name: train_2024-12-01-18-22-24
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# train_2024-12-01-18-22-24
+
+This model is a fine-tuned version of [llava-hf/LLaVA-NeXT-Video-7B-hf](https://huggingface.co/llava-hf/LLaVA-NeXT-Video-7B-hf) on the merger, the LLM_dataset(4o) and the LLM_dataset(4mini) datasets.
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 5e-05
+- train_batch_size: 2
+- eval_batch_size: 8
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 2
+- gradient_accumulation_steps: 8
+- total_train_batch_size: 32
+- total_eval_batch_size: 16
+- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 100
+- num_epochs: 1.0
+
+### Training results
+
+
+
+### Framework versions
+
+- PEFT 0.12.0
+- Transformers 4.46.1
+- Pytorch 2.3.1+cu121
+- Datasets 3.1.0
+- Tokenizers 0.20.3

adapter_config.json

@@ -0,0 +1,26 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "llava-hf/LLaVA-NeXT-Video-7B-hf",
+  "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": 16,
+  "lora_dropout": 0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": "^(?!.*vision_tower).*(?:down_proj|v_proj|k_proj|q_proj|up_proj|o_proj|gate_proj).*",
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16c3e62651b9f8eb60399f5f1140c8b200297419a976c8830d669d09d29a8813
+size 40043208

added_tokens.json

@@ -0,0 +1,4 @@
+{
+  "<image>": 32001,
+  "<video>": 32000
+}

all_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.9832402234636871,
+    "total_flos": 8.043290589292134e+16,
+    "train_loss": 0.8802601207386364,
+    "train_runtime": 618.3986,
+    "train_samples_per_second": 1.155,
+    "train_steps_per_second": 0.036
+}

bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d8326ca28c7c35adb2f9ecc7ecbb84faee0d1d7449e25b8c4ae71aa53ed7da23
+size 119962160

bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd062851371296c614e0bdae1d95959deaaa82e72269ea37d527d89c79951903
+size 119962288

chat_template.json

@@ -0,0 +1,3 @@
+{
+  "chat_template": "{% for message in messages %}{% if message['role'] != 'system' %}{{ message['role'].upper() + ': '}}{% endif %}{# Render all images first #}{% for content in message['content'] | selectattr('type', 'equalto', 'image') %}{{ '<image>\n' }}{% endfor %}{# Render all images first #}{% for content in message['content'] | selectattr('type', 'equalto', 'video') %}{{ '<video>\n' }}{% endfor %}{# Render all text next #}{% if message['role'] != 'assistant' %}{% for content in message['content'] | selectattr('type', 'equalto', 'text') %}{{ content['text'] + ' '}}{% endfor %}{% else %}{% for content in message['content'] | selectattr('type', 'equalto', 'text') %}{% generation %}{{ content['text'] + ' '}}{% endgeneration %}{% endfor %}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ 'ASSISTANT:' }}{% endif %}"
+}

latest

@@ -0,0 +1 @@
+global_step22

llamaboard_config.yaml

@@ -0,0 +1,68 @@
+top.booster: flashattn2
+top.checkpoint_path: []
+top.finetuning_type: lora
+top.model_name: LLaVA-NeXT-Video-7B-Chat
+top.quantization_bit: '4'
+top.quantization_method: bitsandbytes
+top.rope_scaling: none
+top.template: llava_next_video
+train.additional_target: ''
+train.badam_mode: layer
+train.badam_switch_interval: 50
+train.badam_switch_mode: ascending
+train.badam_update_ratio: 0.05
+train.batch_size: 2
+train.compute_type: bf16
+train.create_new_adapter: false
+train.cutoff_len: 4096
+train.dataset:
+- merger
+- LLM_dataset(4o)
+- LLM_dataset(4mini)
+train.dataset_dir: /media/dl/7DC4-B1CE/500_video
+train.ds_offload: false
+train.ds_stage: '2'
+train.extra_args: '{"optim": "adamw_torch"}'
+train.freeze_extra_modules: ''
+train.freeze_trainable_layers: 2
+train.freeze_trainable_modules: all
+train.galore_rank: 16
+train.galore_scale: 0.25
+train.galore_target: all
+train.galore_update_interval: 200
+train.gradient_accumulation_steps: 8
+train.learning_rate: 5e-5
+train.logging_steps: 5
+train.lora_alpha: 16
+train.lora_dropout: 0
+train.lora_rank: 8
+train.lora_target: ''
+train.loraplus_lr_ratio: 0
+train.lr_scheduler_type: cosine
+train.mask_history: false
+train.max_grad_norm: '1.0'
+train.max_samples: '100000'
+train.neat_packing: false
+train.neftune_alpha: 0
+train.num_train_epochs: '1'
+train.packing: false
+train.ppo_score_norm: false
+train.ppo_whiten_rewards: false
+train.pref_beta: 0.1
+train.pref_ftx: 0
+train.pref_loss: sigmoid
+train.report_to: false
+train.resize_vocab: false
+train.reward_model: null
+train.save_steps: 100
+train.shift_attn: false
+train.train_on_prompt: false
+train.training_stage: Supervised Fine-Tuning
+train.use_badam: false
+train.use_dora: false
+train.use_galore: false
+train.use_llama_pro: false
+train.use_pissa: false
+train.use_rslora: false
+train.val_size: 0
+train.warmup_steps: 100

mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2b6318bc36a97cfe9e658ba309603c187ee05cdb5ca8ca7faa8c488c8b3366d
+size 149489983

preprocessor_config.json

@@ -0,0 +1,51 @@
+{
+  "crop_size": {
+    "height": 336,
+    "width": 336
+  },
+  "do_center_crop": true,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_pad": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_grid_pinpoints": [
+    [
+      336,
+      672
+    ],
+    [
+      672,
+      336
+    ],
+    [
+      672,
+      672
+    ],
+    [
+      1008,
+      336
+    ],
+    [
+      336,
+      1008
+    ]
+  ],
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "LlavaNextImageProcessor",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "processor_class": "LlavaNextVideoProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 336
+  }
+}

processor_config.json

@@ -0,0 +1,7 @@
+{
+  "image_token": "<image>",
+  "patch_size": 14,
+  "processor_class": "LlavaNextVideoProcessor",
+  "video_token": "<video>",
+  "vision_feature_select_strategy": "default"
+}

rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c8d6a959372d5e0c2ea025dd26c9d0ad2046fce19352056cae8074dcbd0a6fd4
+size 14512

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f68a37892a1b445d21bb35cc10bf7a058a6f9ec8c363f5ed156ff4f49d90fb6
+size 14512

running_log.txt

@@ -0,0 +1,834 @@
+[INFO|2024-12-01 18:24:23] parser.py:355 >> Process rank: 1, device: cuda:1, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16
+
+[WARNING|2024-12-01 18:24:23] logging.py:162 >> We recommend enable `upcast_layernorm` in quantized training.
+
+[WARNING|2024-12-01 18:24:23] logging.py:162 >> `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training.
+
+[INFO|2024-12-01 18:24:23] parser.py:355 >> Process rank: 0, device: cuda:0, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16
+
+[INFO|2024-12-01 18:24:23] configuration_utils.py:679 >> loading configuration file config.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/config.json
+
+[INFO|2024-12-01 18:24:23] configuration_utils.py:746 >> Model config LlavaNextVideoConfig {
+  "_name_or_path": "llava-hf/LLaVA-NeXT-Video-7B-hf",
+  "architectures": [
+    "LlavaNextVideoForConditionalGeneration"
+  ],
+  "ignore_index": -100,
+  "image_grid_pinpoints": [
+    [
+      336,
+      672
+    ],
+    [
+      672,
+      336
+    ],
+    [
+      672,
+      672
+    ],
+    [
+      1008,
+      336
+    ],
+    [
+      336,
+      1008
+    ]
+  ],
+  "image_seq_length": 576,
+  "image_token_index": 32001,
+  "model_type": "llava_next_video",
+  "projector_hidden_act": "gelu",
+  "spatial_pool_mode": "average",
+  "spatial_pool_out_channels": 1024,
+  "spatial_pool_stride": 2,
+  "text_config": {
+    "_attn_implementation_autoset": false,
+    "_name_or_path": "lmsys/vicuna-7b-v1.5",
+    "add_cross_attention": false,
+    "architectures": [
+      "LlamaForCausalLM"
+    ],
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": 1,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": 2,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "head_dim": 128,
+    "hidden_act": "silu",
+    "hidden_size": 4096,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "intermediate_size": 11008,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "max_position_embeddings": 4096,
+    "min_length": 0,
+    "mlp_bias": false,
+    "model_type": "llama",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 32,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_hidden_layers": 32,
+    "num_key_value_heads": 32,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": 0,
+    "prefix": null,
+    "pretraining_tp": 1,
+    "problem_type": null,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "rms_norm_eps": 1e-05,
+    "rope_scaling": {
+      "factor": 2.5,
+      "rope_type": "linear",
+      "type": "linear"
+    },
+    "rope_theta": 10000.0,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": false,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": "float16",
+    "torchscript": false,
+    "type": "linear",
+    "typical_p": 1.0,
+    "use_bfloat16": false,
+    "use_cache": true,
+    "vocab_size": 32064
+  },
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.46.1",
+  "use_image_newline_parameter": true,
+  "video_seq_length": 288,
+  "video_token_index": 32000,
+  "vision_config": {
+    "_attn_implementation_autoset": false,
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": null,
+    "attention_dropout": 0.0,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": null,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_act": "quick_gelu",
+    "hidden_size": 1024,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "image_size": 336,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "intermediate_size": 4096,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-05,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "clip_vision_model",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 16,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_channels": 3,
+    "num_hidden_layers": 24,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": null,
+    "patch_size": 14,
+    "prefix": null,
+    "problem_type": null,
+    "projection_dim": 768,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": null,
+    "torchscript": false,
+    "typical_p": 1.0,
+    "use_bfloat16": false,
+    "vocab_size": 32000
+  },
+  "vision_feature_layer": -2,
+  "vision_feature_select_strategy": "default"
+}
+
+
+[INFO|2024-12-01 18:24:23] tokenization_utils_base.py:2211 >> loading file tokenizer.model from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/tokenizer.model
+
+[INFO|2024-12-01 18:24:23] tokenization_utils_base.py:2211 >> loading file tokenizer.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/tokenizer.json
+
+[INFO|2024-12-01 18:24:23] tokenization_utils_base.py:2211 >> loading file added_tokens.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/added_tokens.json
+
+[INFO|2024-12-01 18:24:23] tokenization_utils_base.py:2211 >> loading file special_tokens_map.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/special_tokens_map.json
+
+[INFO|2024-12-01 18:24:23] tokenization_utils_base.py:2211 >> loading file tokenizer_config.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/tokenizer_config.json
+
+[INFO|2024-12-01 18:24:23] tokenization_utils_base.py:2475 >> Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+
+[INFO|2024-12-01 18:24:24] processing_utils.py:695 >> loading configuration file processor_config.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/processor_config.json
+
+[INFO|2024-12-01 18:24:24] image_processing_base.py:375 >> loading configuration file preprocessor_config.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/preprocessor_config.json
+
+[INFO|2024-12-01 18:24:24] image_processing_base.py:429 >> Image processor LlavaNextVideoImageProcessor {
+  "crop_size": {
+    "height": 336,
+    "width": 336
+  },
+  "do_center_crop": true,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_pad": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_grid_pinpoints": [
+    [
+      336,
+      672
+    ],
+    [
+      672,
+      336
+    ],
+    [
+      672,
+      672
+    ],
+    [
+      1008,
+      336
+    ],
+    [
+      336,
+      1008
+    ]
+  ],
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "LlavaNextVideoImageProcessor",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "processor_class": "LlavaNextVideoProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 336
+  }
+}
+
+
+[INFO|2024-12-01 18:24:24] image_processing_base.py:375 >> loading configuration file preprocessor_config.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/preprocessor_config.json
+
+[INFO|2024-12-01 18:24:24] image_processing_base.py:429 >> Image processor LlavaNextImageProcessor {
+  "crop_size": {
+    "height": 336,
+    "width": 336
+  },
+  "do_center_crop": true,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_pad": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_grid_pinpoints": [
+    [
+      336,
+      672
+    ],
+    [
+      672,
+      336
+    ],
+    [
+      672,
+      672
+    ],
+    [
+      1008,
+      336
+    ],
+    [
+      336,
+      1008
+    ]
+  ],
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "LlavaNextImageProcessor",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "processor_class": "LlavaNextVideoProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 336
+  }
+}
+
+
+[INFO|2024-12-01 18:24:25] tokenization_utils_base.py:2211 >> loading file tokenizer.model from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/tokenizer.model
+
+[INFO|2024-12-01 18:24:25] tokenization_utils_base.py:2211 >> loading file tokenizer.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/tokenizer.json
+
+[INFO|2024-12-01 18:24:25] tokenization_utils_base.py:2211 >> loading file added_tokens.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/added_tokens.json
+
+[INFO|2024-12-01 18:24:25] tokenization_utils_base.py:2211 >> loading file special_tokens_map.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/special_tokens_map.json
+
+[INFO|2024-12-01 18:24:25] tokenization_utils_base.py:2211 >> loading file tokenizer_config.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/tokenizer_config.json
+
+[INFO|2024-12-01 18:24:25] tokenization_utils_base.py:2475 >> Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+
+[INFO|2024-12-01 18:24:25] processing_utils.py:695 >> loading configuration file processor_config.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/processor_config.json
+
+[WARNING|2024-12-01 18:24:25] processing_utils.py:1005 >> Some kwargs in processor config are unused and will not have any effect: num_additional_image_tokens. 
+
+[INFO|2024-12-01 18:24:25] processing_utils.py:755 >> Processor LlavaNextVideoProcessor:
+- video_processor: LlavaNextVideoImageProcessor {
+  "crop_size": {
+    "height": 336,
+    "width": 336
+  },
+  "do_center_crop": true,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_pad": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_grid_pinpoints": [
+    [
+      336,
+      672
+    ],
+    [
+      672,
+      336
+    ],
+    [
+      672,
+      672
+    ],
+    [
+      1008,
+      336
+    ],
+    [
+      336,
+      1008
+    ]
+  ],
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "LlavaNextVideoImageProcessor",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "processor_class": "LlavaNextVideoProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 336
+  }
+}
+
+- image_processor: LlavaNextImageProcessor {
+  "crop_size": {
+    "height": 336,
+    "width": 336
+  },
+  "do_center_crop": true,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_pad": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_grid_pinpoints": [
+    [
+      336,
+      672
+    ],
+    [
+      672,
+      336
+    ],
+    [
+      672,
+      672
+    ],
+    [
+      1008,
+      336
+    ],
+    [
+      336,
+      1008
+    ]
+  ],
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "LlavaNextImageProcessor",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "processor_class": "LlavaNextVideoProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 336
+  }
+}
+
+- tokenizer: LlamaTokenizerFast(name_or_path='llava-hf/LLaVA-NeXT-Video-7B-hf', vocab_size=32000, model_max_length=4096, is_fast=True, padding_side='left', truncation_side='right', special_tokens={'bos_token': '<s>', 'eos_token': '</s>', 'unk_token': '<unk>', 'pad_token': '<unk>'}, clean_up_tokenization_spaces=False),  added_tokens_decoder={
+	0: AddedToken("<unk>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
+	1: AddedToken("<s>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
+	2: AddedToken("</s>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
+	32000: AddedToken("<video>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
+	32001: AddedToken("<image>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
+}
+
+{
+  "image_token": "<image>",
+  "patch_size": 14,
+  "processor_class": "LlavaNextVideoProcessor",
+  "video_token": "<video>",
+  "vision_feature_select_strategy": "default"
+}
+
+
+[INFO|2024-12-01 18:24:25] logging.py:157 >> Loading dataset merger500.json...
+
+[WARNING|2024-12-01 18:24:25] processing_utils.py:1005 >> Some kwargs in processor config are unused and will not have any effect: num_additional_image_tokens. 
+
+[INFO|2024-12-01 18:24:27] logging.py:157 >> Loading dataset LLM_dataset(4o).json...
+
+[INFO|2024-12-01 18:24:27] logging.py:157 >> Loading dataset LLM_dataset(4mini).json...
+
+[INFO|2024-12-01 18:25:48] configuration_utils.py:679 >> loading configuration file config.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/config.json
+
+[INFO|2024-12-01 18:25:48] configuration_utils.py:746 >> Model config LlavaNextVideoConfig {
+  "_name_or_path": "llava-hf/LLaVA-NeXT-Video-7B-hf",
+  "architectures": [
+    "LlavaNextVideoForConditionalGeneration"
+  ],
+  "ignore_index": -100,
+  "image_grid_pinpoints": [
+    [
+      336,
+      672
+    ],
+    [
+      672,
+      336
+    ],
+    [
+      672,
+      672
+    ],
+    [
+      1008,
+      336
+    ],
+    [
+      336,
+      1008
+    ]
+  ],
+  "image_seq_length": 576,
+  "image_token_index": 32001,
+  "model_type": "llava_next_video",
+  "projector_hidden_act": "gelu",
+  "spatial_pool_mode": "average",
+  "spatial_pool_out_channels": 1024,
+  "spatial_pool_stride": 2,
+  "text_config": {
+    "_attn_implementation_autoset": false,
+    "_name_or_path": "lmsys/vicuna-7b-v1.5",
+    "add_cross_attention": false,
+    "architectures": [
+      "LlamaForCausalLM"
+    ],
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": 1,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": 2,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "head_dim": 128,
+    "hidden_act": "silu",
+    "hidden_size": 4096,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "intermediate_size": 11008,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "max_position_embeddings": 4096,
+    "min_length": 0,
+    "mlp_bias": false,
+    "model_type": "llama",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 32,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_hidden_layers": 32,
+    "num_key_value_heads": 32,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": 0,
+    "prefix": null,
+    "pretraining_tp": 1,
+    "problem_type": null,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "rms_norm_eps": 1e-05,
+    "rope_scaling": {
+      "factor": 2.5,
+      "rope_type": "linear",
+      "type": "linear"
+    },
+    "rope_theta": 10000.0,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": false,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": "float16",
+    "torchscript": false,
+    "type": "linear",
+    "typical_p": 1.0,
+    "use_bfloat16": false,
+    "use_cache": true,
+    "vocab_size": 32064
+  },
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.46.1",
+  "use_image_newline_parameter": true,
+  "video_seq_length": 288,
+  "video_token_index": 32000,
+  "vision_config": {
+    "_attn_implementation_autoset": false,
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": null,
+    "attention_dropout": 0.0,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": null,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_act": "quick_gelu",
+    "hidden_size": 1024,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "image_size": 336,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "intermediate_size": 4096,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-05,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "clip_vision_model",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 16,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_channels": 3,
+    "num_hidden_layers": 24,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": null,
+    "patch_size": 14,
+    "prefix": null,
+    "problem_type": null,
+    "projection_dim": 768,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": null,
+    "torchscript": false,
+    "typical_p": 1.0,
+    "use_bfloat16": false,
+    "vocab_size": 32000
+  },
+  "vision_feature_layer": -2,
+  "vision_feature_select_strategy": "default"
+}
+
+
+[INFO|2024-12-01 18:25:48] logging.py:157 >> Quantizing model to 4 bit with bitsandbytes.
+
+[INFO|2024-12-01 18:25:48] modeling_utils.py:3937 >> loading weights file model.safetensors from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/model.safetensors.index.json
+
+[INFO|2024-12-01 18:25:48] modeling_utils.py:1670 >> Instantiating LlavaNextVideoForConditionalGeneration model under default dtype torch.bfloat16.
+
+[INFO|2024-12-01 18:25:48] configuration_utils.py:1096 >> Generate config GenerationConfig {}
+
+
+[INFO|2024-12-01 18:25:48] modeling_utils.py:1670 >> Instantiating CLIPVisionModel model under default dtype torch.bfloat16.
+
+[INFO|2024-12-01 18:25:48] modeling_utils.py:1670 >> Instantiating LlamaForCausalLM model under default dtype torch.bfloat16.
+
+[INFO|2024-12-01 18:25:48] configuration_utils.py:1096 >> Generate config GenerationConfig {
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "pad_token_id": 0
+}
+
+
+[INFO|2024-12-01 18:25:52] modeling_utils.py:4800 >> All model checkpoint weights were used when initializing LlavaNextVideoForConditionalGeneration.
+
+
+[INFO|2024-12-01 18:25:52] modeling_utils.py:4808 >> All the weights of LlavaNextVideoForConditionalGeneration were initialized from the model checkpoint at llava-hf/LLaVA-NeXT-Video-7B-hf.
+If your task is similar to the task the model of the checkpoint was trained on, you can already use LlavaNextVideoForConditionalGeneration for predictions without further training.
+
+[INFO|2024-12-01 18:25:52] configuration_utils.py:1051 >> loading configuration file generation_config.json from cache at /home/dl/.cache/huggingface/hub/models--llava-hf--LLaVA-NeXT-Video-7B-hf/snapshots/b3b624d0915bb487ef1abb15255aaa2cd5581205/generation_config.json
+
+[INFO|2024-12-01 18:25:52] configuration_utils.py:1096 >> Generate config GenerationConfig {
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "pad_token_id": 0
+}
+
+
+[INFO|2024-12-01 18:25:53] logging.py:157 >> Gradient checkpointing enabled.
+
+[INFO|2024-12-01 18:25:53] logging.py:157 >> Casting multimodal projector outputs in torch.bfloat16.
+
+[INFO|2024-12-01 18:25:53] logging.py:157 >> Using FlashAttention-2 for faster training and inference.
+
+[INFO|2024-12-01 18:25:53] logging.py:157 >> Upcasting trainable params to float32.
+
+[INFO|2024-12-01 18:25:53] logging.py:157 >> Fine-tuning method: LoRA
+
+[INFO|2024-12-01 18:25:53] logging.py:157 >> Found linear modules: down_proj,v_proj,k_proj,q_proj,up_proj,o_proj,gate_proj
+
+[INFO|2024-12-01 18:25:53] logging.py:157 >> trainable params: 19,988,480 || all params: 7,083,419,648 || trainable%: 0.2822
+
+[INFO|2024-12-01 18:25:53] trainer.py:698 >> Using auto half precision backend
+
+[INFO|2024-12-01 18:25:55] trainer.py:2313 >> ***** Running training *****
+
+[INFO|2024-12-01 18:25:55] trainer.py:2314 >>   Num examples = 714
+
+[INFO|2024-12-01 18:25:55] trainer.py:2315 >>   Num Epochs = 1
+
+[INFO|2024-12-01 18:25:55] trainer.py:2316 >>   Instantaneous batch size per device = 2
+
+[INFO|2024-12-01 18:25:55] trainer.py:2319 >>   Total train batch size (w. parallel, distributed & accumulation) = 32
+
+[INFO|2024-12-01 18:25:55] trainer.py:2320 >>   Gradient Accumulation steps = 8
+
+[INFO|2024-12-01 18:25:55] trainer.py:2321 >>   Total optimization steps = 22
+
+[INFO|2024-12-01 18:25:55] trainer.py:2322 >>   Number of trainable parameters = 19,988,480
+
+[WARNING|2024-12-01 18:25:59] logging.py:168 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`.
+
+[INFO|2024-12-01 18:28:23] logging.py:157 >> {'loss': 0.8615, 'learning_rate': 2.5000e-06, 'epoch': 0.22}
+
+[INFO|2024-12-01 18:30:41] logging.py:157 >> {'loss': 0.8940, 'learning_rate': 5.0000e-06, 'epoch': 0.45}
+
+[INFO|2024-12-01 18:33:04] logging.py:157 >> {'loss': 0.8808, 'learning_rate': 7.5000e-06, 'epoch': 0.67}
+
+[INFO|2024-12-01 18:35:22] logging.py:157 >> {'loss': 0.8746, 'learning_rate': 1.0000e-05, 'epoch': 0.89}
+
+[INFO|2024-12-01 18:36:12] trainer.py:3801 >> Saving model checkpoint to saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/checkpoint-22
+
+[INFO|2024-12-01 18:36:13] tokenization_utils_base.py:2646 >> tokenizer config file saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/checkpoint-22/tokenizer_config.json
+
+[INFO|2024-12-01 18:36:13] tokenization_utils_base.py:2655 >> Special tokens file saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/checkpoint-22/special_tokens_map.json
+
+[INFO|2024-12-01 18:36:13] image_processing_base.py:258 >> Image processor saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/checkpoint-22/preprocessor_config.json
+
+[INFO|2024-12-01 18:36:13] image_processing_base.py:258 >> Image processor saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/checkpoint-22/preprocessor_config.json
+
+[INFO|2024-12-01 18:36:13] tokenization_utils_base.py:2646 >> tokenizer config file saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/checkpoint-22/tokenizer_config.json
+
+[INFO|2024-12-01 18:36:13] tokenization_utils_base.py:2655 >> Special tokens file saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/checkpoint-22/special_tokens_map.json
+
+[INFO|2024-12-01 18:36:13] processing_utils.py:541 >> chat template saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/checkpoint-22/chat_template.json
+
+[INFO|2024-12-01 18:36:13] processing_utils.py:547 >> processor saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/checkpoint-22/processor_config.json
+
+[INFO|2024-12-01 18:36:13] trainer.py:2584 >> 
+
+Training completed. Do not forget to share your model on huggingface.co/models =)
+
+
+
+[INFO|2024-12-01 18:36:13] image_processing_base.py:258 >> Image processor saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/preprocessor_config.json
+
+[INFO|2024-12-01 18:36:13] image_processing_base.py:258 >> Image processor saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/preprocessor_config.json
+
+[INFO|2024-12-01 18:36:13] tokenization_utils_base.py:2646 >> tokenizer config file saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/tokenizer_config.json
+
+[INFO|2024-12-01 18:36:13] tokenization_utils_base.py:2655 >> Special tokens file saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/special_tokens_map.json
+
+[INFO|2024-12-01 18:36:14] processing_utils.py:541 >> chat template saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/chat_template.json
+
+[INFO|2024-12-01 18:36:14] processing_utils.py:547 >> processor saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/processor_config.json
+
+[INFO|2024-12-01 18:36:14] trainer.py:3801 >> Saving model checkpoint to saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24
+
+[INFO|2024-12-01 18:36:15] tokenization_utils_base.py:2646 >> tokenizer config file saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/tokenizer_config.json
+
+[INFO|2024-12-01 18:36:15] tokenization_utils_base.py:2655 >> Special tokens file saved in saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24/special_tokens_map.json
+
+[WARNING|2024-12-01 18:36:15] logging.py:162 >> No metric eval_loss to plot.
+
+[WARNING|2024-12-01 18:36:15] logging.py:162 >> No metric eval_accuracy to plot.
+
+[INFO|2024-12-01 18:36:15] modelcard.py:449 >> Dropping the following result as it does not have all the necessary fields:
+{'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}}
+

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab66602d7e6c10eee8866ffc8c6b4541b799bb1c62eb8be9713bd239d7fe2942
+size 1064

special_tokens_map.json

@@ -0,0 +1,30 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

tokenizer_config.json

@@ -0,0 +1,68 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "add_prefix_space": true,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "32000": {
+      "content": "<video>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "32001": {
+      "content": "<image>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "chat_template": "{% set system_message = 'A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user\\'s questions.' %}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% endif %}{% if system_message is defined %}{{ system_message }}{% endif %}{% for message in loop_messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ 'USER: ' + content + ' ASSISTANT:' }}{% elif message['role'] == 'assistant' %}{{ content + '</s>' }}{% endif %}{% endfor %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "extra_special_tokens": {
+    "image_token": "<image>",
+    "video_token": "<video>"
+  },
+  "image_token": "<image>",
+  "legacy": false,
+  "model_max_length": 4096,
+  "pad_token": "<unk>",
+  "padding_side": "right",
+  "processor_class": "LlavaNextVideoProcessor",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "split_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false,
+  "video_token": "<video>"
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.9832402234636871,
+    "total_flos": 8.043290589292134e+16,
+    "train_loss": 0.8802601207386364,
+    "train_runtime": 618.3986,
+    "train_samples_per_second": 1.155,
+    "train_steps_per_second": 0.036
+}

trainer_log.jsonl

@@ -0,0 +1,5 @@
+{"current_steps": 5, "total_steps": 22, "loss": 0.8615, "lr": 2.5e-06, "epoch": 0.22346368715083798, "percentage": 22.73, "elapsed_time": "0:02:28", "remaining_time": "0:08:24"}
+{"current_steps": 10, "total_steps": 22, "loss": 0.894, "lr": 5e-06, "epoch": 0.44692737430167595, "percentage": 45.45, "elapsed_time": "0:04:46", "remaining_time": "0:05:43"}
+{"current_steps": 15, "total_steps": 22, "loss": 0.8808, "lr": 7.5e-06, "epoch": 0.6703910614525139, "percentage": 68.18, "elapsed_time": "0:07:08", "remaining_time": "0:03:20"}
+{"current_steps": 20, "total_steps": 22, "loss": 0.8746, "lr": 1e-05, "epoch": 0.8938547486033519, "percentage": 90.91, "elapsed_time": "0:09:27", "remaining_time": "0:00:56"}
+{"current_steps": 22, "total_steps": 22, "epoch": 0.9832402234636871, "percentage": 100.0, "elapsed_time": "0:10:18", "remaining_time": "0:00:00"}

trainer_state.json

@@ -0,0 +1,61 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9832402234636871,
+  "eval_steps": 500,
+  "global_step": 22,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.22346368715083798,
+      "grad_norm": 0.4495355188846588,
+      "learning_rate": 2.5e-06,
+      "loss": 0.8615,
+      "step": 5
+    },
+    {
+      "epoch": 0.44692737430167595,
+      "grad_norm": 0.49592095613479614,
+      "learning_rate": 5e-06,
+      "loss": 0.894,
+      "step": 10
+    },
+    {
+      "epoch": 0.6703910614525139,
+      "grad_norm": 0.47344380617141724,
+      "learning_rate": 7.5e-06,
+      "loss": 0.8808,
+      "step": 15
+    },
+    {
+      "epoch": 0.8938547486033519,
+      "grad_norm": 0.5614600777626038,
+      "learning_rate": 1e-05,
+      "loss": 0.8746,
+      "step": 20
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 22,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 8.043290589292134e+16,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4458e5e72517233363af05f835b35082429196f135ba58f1802787df1f5a0da
+size 7032

training_args.yaml

@@ -0,0 +1,34 @@
+bf16: true
+cutoff_len: 4096
+dataset: merger,LLM_dataset(4o),LLM_dataset(4mini)
+dataset_dir: /media/dl/7DC4-B1CE/500_video
+ddp_timeout: 180000000
+deepspeed: cache/ds_z2_config.json
+do_train: true
+finetuning_type: lora
+flash_attn: fa2
+gradient_accumulation_steps: 8
+learning_rate: 5.0e-05
+logging_steps: 5
+lora_alpha: 16
+lora_dropout: 0
+lora_rank: 8
+lora_target: all
+lr_scheduler_type: cosine
+max_grad_norm: 1.0
+max_samples: 100000
+model_name_or_path: llava-hf/LLaVA-NeXT-Video-7B-hf
+num_train_epochs: 1.0
+optim: adamw_torch
+output_dir: saves/LLaVA-NeXT-Video-7B-Chat/lora/train_2024-12-01-18-22-24
+packing: false
+per_device_train_batch_size: 2
+plot_loss: true
+preprocessing_num_workers: 16
+quantization_bit: 4
+quantization_method: bitsandbytes
+report_to: none
+save_steps: 100
+stage: sft
+template: llava_next_video
+warmup_steps: 100

zero_to_fp32.py

@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)