Add functionality to download models from sources other than HuggingF…

…ace (#946)

support openmind model and dataset

docs/zh_cn/training/modify_settings.rst

@@ -252,6 +252,22 @@ XTuner 使用 MMEngine 的「纯 Python 风格的配置文件」，直接利用
 1.  参考 `文档 <../preparation/pretrained_model.md>`__ 将其下载至本地
 2.  修改\ ``pretrained_model_name_or_path``\ 。
 
+使用 openMind 模型？
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+可在配置文件中新增 ``model_resource`` 参数， ``args`` 用作可变参数（如下载私有模型需传入token的情况）：
+
+.. code:: python
+   from openmind_hub import snapshot_download 
+
+   # Model
+   pretrained_model_name_or_path = 'Tianjin_Ascend/Qwen1.5-4B'
+   model_resource = {
+      "fn": snapshot_download,
+      "args":{ 
+         # "token":"xxxxxxxxxx"
+      }
+   }
+
 微调类型
 -------------
 

docs/zh_cn/training/open_source_dataset.rst

@@ -31,6 +31,8 @@ XTuner 使用上游库 ``datasets`` 的统一载入接口 ``load_dataset``\ 。
     ``dataset=dict(type=load_dataset, path=data_path)`` 中的 ``path``
     参数即可。
 
+    若想使用 openMind 数据集，可将 ``dataset=dict(type=load_dataset, path=data_path)`` 中的 ``type`` 替换为 ``openmind.OmDataset``。
+
 
 字段格式
 --------

xtuner/configs/qwen/qwen1_5/qwen1_5_4b/qwen1_5_4b_qlora_alpaca_e3_openmind.py

@@ -0,0 +1,230 @@
+# Copyright (c) OpenMMLab. All rights reserved.
+import torch
+from datasets import load_dataset
+from mmengine.dataset import DefaultSampler
+from mmengine.hooks import (CheckpointHook, DistSamplerSeedHook, IterTimerHook,
+                            LoggerHook, ParamSchedulerHook)
+from mmengine.optim import AmpOptimWrapper, CosineAnnealingLR, LinearLR
+from peft import LoraConfig
+from torch.optim import AdamW
+from transformers import (AutoModelForCausalLM, AutoTokenizer,
+                          BitsAndBytesConfig)
+
+from xtuner.dataset import process_hf_dataset
+from xtuner.dataset.collate_fns import default_collate_fn
+from xtuner.dataset.map_fns import alpaca_map_fn, template_map_fn_factory
+from xtuner.engine.hooks import (DatasetInfoHook, EvaluateChatHook,
+                                 VarlenAttnArgsToMessageHubHook)
+from xtuner.engine.runner import TrainLoop
+from xtuner.model import SupervisedFinetune
+from xtuner.parallel.sequence import SequenceParallelSampler
+from xtuner.utils import PROMPT_TEMPLATE, SYSTEM_TEMPLATE
+from openmind_hub import snapshot_download
+from openmind import OmDataset
+
+#######################################################################
+#                          PART 1  Settings                           #
+#######################################################################
+# Model
+pretrained_model_name_or_path = 'Tianjin_Ascend/Qwen1.5-4B'
+model_resource = {
+    "fn": snapshot_download,
+    "args":{ 
+        # "token":"xxxxxxxxxx"
+    }
+}
+use_varlen_attn = False
+
+# Data
+alpaca_en_path = 'AI_Connect/alpaca'
+prompt_template = PROMPT_TEMPLATE.default
+max_length = 2048
+pack_to_max_length = True
+
+# parallel
+sequence_parallel_size = 1
+
+# Scheduler & Optimizer
+batch_size = 1  # per_device
+accumulative_counts = 16
+accumulative_counts *= sequence_parallel_size
+dataloader_num_workers = 0
+max_epochs = 3
+optim_type = AdamW
+lr = 2e-4
+betas = (0.9, 0.999)
+weight_decay = 0
+max_norm = 1  # grad clip
+warmup_ratio = 0.03
+
+# Save
+save_steps = 500
+save_total_limit = 2  # Maximum checkpoints to keep (-1 means unlimited)
+
+# Evaluate the generation performance during the training
+evaluation_freq = 500
+SYSTEM = SYSTEM_TEMPLATE.alpaca
+evaluation_inputs = [
+    '请给我介绍五个上海的景点', 'Please tell me five scenic spots in Shanghai'
+]
+
+#######################################################################
+#                      PART 2  Model & Tokenizer                      #
+#######################################################################
+tokenizer = dict(
+    type=AutoTokenizer.from_pretrained,
+    pretrained_model_name_or_path=pretrained_model_name_or_path,
+    trust_remote_code=True,
+    padding_side='right')
+
+model = dict(
+    type=SupervisedFinetune,
+    use_varlen_attn=use_varlen_attn,
+    llm=dict(
+        type=AutoModelForCausalLM.from_pretrained,
+        pretrained_model_name_or_path=pretrained_model_name_or_path,
+        trust_remote_code=True,
+        torch_dtype=torch.float16,
+        # NPU does not support quantization 
+        # quantization_config=dict(
+        #     type=BitsAndBytesConfig,
+        #     load_in_4bit=True,
+        #     load_in_8bit=False,
+        #     llm_int8_threshold=6.0,
+        #     llm_int8_has_fp16_weight=False,
+        #     bnb_4bit_compute_dtype=torch.float16,
+        #     bnb_4bit_use_double_quant=True,
+        #     bnb_4bit_quant_type='nf4')
+            ),
+    lora=dict(
+        type=LoraConfig,
+        r=64,
+        lora_alpha=16,
+        lora_dropout=0.1,
+        bias='none',
+        task_type='CAUSAL_LM'))
+
+#######################################################################
+#                      PART 3  Dataset & Dataloader                   #
+#######################################################################
+alpaca_en = dict(
+    type=process_hf_dataset,
+    dataset=dict(type=OmDataset.load_dataset, path=alpaca_en_path),
+    tokenizer=tokenizer,
+    max_length=max_length,
+    dataset_map_fn=alpaca_map_fn,
+    template_map_fn=dict(
+        type=template_map_fn_factory, template=prompt_template),
+    remove_unused_columns=True,
+    shuffle_before_pack=True,
+    pack_to_max_length=pack_to_max_length,
+    use_varlen_attn=use_varlen_attn)
+
+sampler = SequenceParallelSampler \
+    if sequence_parallel_size > 1 else DefaultSampler
+
+train_dataloader = dict(
+    batch_size=batch_size,
+    num_workers=dataloader_num_workers,
+    dataset=alpaca_en,
+    sampler=dict(type=sampler, shuffle=True),
+    collate_fn=dict(type=default_collate_fn, use_varlen_attn=use_varlen_attn))
+
+#######################################################################
+#                    PART 4  Scheduler & Optimizer                    #
+#######################################################################
+# optimizer
+optim_wrapper = dict(
+    type=AmpOptimWrapper,
+    optimizer=dict(
+        type=optim_type, lr=lr, betas=betas, weight_decay=weight_decay),
+    clip_grad=dict(max_norm=max_norm, error_if_nonfinite=False),
+    accumulative_counts=accumulative_counts,
+    loss_scale='dynamic',
+    dtype='float16')
+
+# learning policy
+# More information: https://github.com/open-mmlab/mmengine/blob/main/docs/en/tutorials/param_scheduler.md  # noqa: E501
+param_scheduler = [
+    dict(
+        type=LinearLR,
+        start_factor=1e-5,
+        by_epoch=True,
+        begin=0,
+        end=warmup_ratio * max_epochs,
+        convert_to_iter_based=True),
+    dict(
+        type=CosineAnnealingLR,
+        eta_min=0.0,
+        by_epoch=True,
+        begin=warmup_ratio * max_epochs,
+        end=max_epochs,
+        convert_to_iter_based=True)
+]
+
+# train, val, test setting
+train_cfg = dict(type=TrainLoop, max_epochs=max_epochs)
+
+#######################################################################
+#                           PART 5  Runtime                           #
+#######################################################################
+# Log the dialogue periodically during the training process, optional
+custom_hooks = [
+    dict(type=DatasetInfoHook, tokenizer=tokenizer),
+    dict(
+        type=EvaluateChatHook,
+        tokenizer=tokenizer,
+        every_n_iters=evaluation_freq,
+        evaluation_inputs=evaluation_inputs,
+        system=SYSTEM,
+        prompt_template=prompt_template)
+]
+
+if use_varlen_attn:
+    custom_hooks += [dict(type=VarlenAttnArgsToMessageHubHook)]
+
+# configure default hooks
+default_hooks = dict(
+    # record the time of every iteration.
+    timer=dict(type=IterTimerHook),
+    # print log every 10 iterations.
+    logger=dict(type=LoggerHook, log_metric_by_epoch=False, interval=10),
+    # enable the parameter scheduler.
+    param_scheduler=dict(type=ParamSchedulerHook),
+    # save checkpoint per `save_steps`.
+    checkpoint=dict(
+        type=CheckpointHook,
+        by_epoch=False,
+        interval=save_steps,
+        max_keep_ckpts=save_total_limit),
+    # set sampler seed in distributed evrionment.
+    sampler_seed=dict(type=DistSamplerSeedHook),
+)
+
+# configure environment
+env_cfg = dict(
+    # whether to enable cudnn benchmark
+    cudnn_benchmark=False,
+    # set multi process parameters
+    mp_cfg=dict(mp_start_method='fork', opencv_num_threads=0),
+    # set distributed parameters
+    dist_cfg=dict(backend='nccl'),
+)
+
+# set visualizer
+visualizer = None
+
+# set log level
+log_level = 'INFO'
+
+# load from which checkpoint
+load_from = None
+
+# whether to resume training from the loaded checkpoint
+resume = False
+
+# Defaults to use random seed and disable `deterministic`
+randomness = dict(seed=None, deterministic=False)
+
+# set log processor
+log_processor = dict(by_epoch=False)

xtuner/dataset/__init__.py

@@ -21,7 +21,7 @@
 __all__ = [
     'process_hf_dataset', 'ConcatDataset', 'MOSSSFTDataset',
     'process_ms_dataset', 'LLaVADataset', 'expand2square',
-    'decode_base64_to_image', 'load_image', 'process_ms_dataset',
+    'decode_base64_to_image', 'load_image', 
     'load_intern_repo_tokenized_dataset',
     'load_intern_repo_untokenized_dataset', 'build_packed_dataset',
     'RefCOCOJsonDataset', 'RefCOCOJsonEvalDataset', 'InvRefCOCOJsonDataset',

xtuner/tools/train.py

@@ -23,7 +23,7 @@
 from xtuner.model.utils import LoadWoInit, find_all_linear_names, traverse_dict
 from xtuner.registry import BUILDER, MAP_FUNC
 from xtuner.tools.utils import (auto_dtype_of_deepspeed_config,
-                                get_seed_from_checkpoint)
+                                get_seed_from_checkpoint, set_model_resource)
 
 
 def parse_args():
@@ -124,6 +124,9 @@ def check_cfg(cfg, args):
              'deepspeed_zero3)`.')
 
 
+
+
+
 def main():
     args = parse_args()
 
@@ -136,6 +139,7 @@ def main():
 
     # load config
     cfg = Config.fromfile(args.config)
+    set_model_resource(cfg)
 
     if args.cfg_options is not None:
         cfg.merge_from_dict(args.cfg_options)

xtuner/tools/utils.py

@@ -37,6 +37,24 @@ def get_streamer(model):
     else:
         return DecodeOutputStreamer
 
+def set_model_resource(cfg):
+    if cfg.get("model_resource"):
+        fn = cfg["model_resource"].get("fn")
+        args = cfg["model_resource"].get("args", {})
+        local_path = fn(cfg["pretrained_model_name_or_path"], **args)
+        s = [(cfg._cfg_dict, k, v) for k, v in cfg._cfg_dict.items()]
+        while s:
+            current_d, current_k, current_v = s.pop()
+            if current_k == "pretrained_model_name_or_path":
+                current_d[current_k] = local_path
+
+            if isinstance(current_v, dict):
+                s.extend([(current_v, k, v) for k, v in current_v.items()])
+            elif isinstance(current_v, list):
+                for i in current_v:
+                    if isinstance(i, dict):
+                        s.extend((i, k, v) for k, v in i.items())
+
 
 class DecodeOutputStreamer(BaseStreamer):
     """Default streamer for HuggingFace models."""