git_6663_introduction

README.md

@@ -1 +1,92 @@
-# tutorial
+# 书生大模型实战营（第四期）
+
+![icamp4](https://github.com/user-attachments/assets/cf5eca85-bb13-4b96-aa54-b29fa40d36b7)
+
+
+闯关手册：https://aicarrier.feishu.cn/wiki/QtJnweAW1iFl8LkoMKGcsUS9nld
+
+
+## 1. 关卡
+
+![update](https://github.com/user-attachments/assets/1f65a77a-3490-47ff-b408-be6d7a190802)
+
+
+### 1.1. 入门关卡
+
+||关卡名称|资料|闯关激励|
+|:-----|:----|:----|:-----|
+|第 1 关| Linux 前置基础 |[任务](docs/L0/linux/task.md)、[文档](docs/L0/linux)、[视频](https://www.bilibili.com/video/BV13U1VYmEUr)| 50元算力点 |
+|第 2 关|Python 前置基础 | [任务](docs/L0/Python/task.md)、[文档](docs/L0/Python)、[视频](https://www.bilibili.com/video/BV1u61jYSExg)| 50元算力点|
+|第 3 关|Git 前置基础|[任务](docs/L0/git/task.md)、[文档](docs/L0/git/)、[视频](https://www.bilibili.com/video/BV15MShYkEgg)| 50元算力点 |
+|第 4 关|玩转「HF/魔搭/魔乐」等平台|[任务](docs/L0/maas/task.md)、[文档](docs/L0/maas)、[视频](https://www.bilibili.com/video/BV1XxStYYEH1/)| 50元算力点 |
+
+
+### 1.2. 基础关卡
+
+
+||关卡名称|资料|闯关激励|
+|:-----|:----|:----|:-----|
+|第 1 关| 书生大模型全链路开源体系 |[任务](docs/L1/ToolChain)、[视频](https://www.bilibili.com/video/BV1CkSUYGE1v/)| 100元算力点 |
+|第 2 关| 玩转书生「多模态对话」和「AI搜索」产品 | [任务](docs/L1/InternIntro/tasks.md)、[文档](docs/L1/InternIntro)、[视频](https://www.bilibili.com/video/BV1ExDQYyEAA)| 100元算力点 |
+|第 3 关| 浦语提示词工程实践 | [任务](docs/L1/Prompt/tasks.md)、[文档](docs/L1/Prompt)、[视频](https://www.bilibili.com/video/BV1tjS7YfEWJ/)| 100元算力点 |
+|第 4 关| InternLM + LlamaIndex RAG 实践|[任务](docs/L1/LlamaIndex/task.md)、[文档](docs/L1/LlamaIndex)、[视频](https://www.bilibili.com/video/BV1YzDJY1E2i/)| 100元算力点 |
+|第 5 关| XTuner 微调个人小助手认知 | [任务](docs/L1/XTuner/task.md)、[文档](docs/L1/XTuner)、[视频](https://www.bilibili.com/video/BV1G9SJYGEtD)| 100元算力点 |
+|第 6 关| OpenCompass 评测书生大模型实践 | [任务](docs/L1/Evaluation/task.md)、[文档](docs/L1/Evaluation/)、[视频](https://www.bilibili.com/video/BV1dtD4YKENj)| 100元算力点 |
+
+
+
+### 1.3. 进阶关卡
+
+||关卡名称|资料|闯关激励|
+|:-----|:----|:----|:-----|
+|第 1 关| 探索书生大模型能力边界 | [任务](https://aicarrier.feishu.cn/wiki/SmUNwgM8vizLPRkhqr1cY1BKngb)| 100元算力点 |
+|第 2 关| Lagent：从零搭建你的 Multi-Agent | [任务](docs/L2/Agent/task.md)、[文档](docs/L2/Agent)、[视频](https://www.bilibili.com/video/BV19RzcYaEFy)| 100元算力点 |
+|第 3 关| LMDeploy 量化部署进阶实践 | [任务](docs/L2/LMDeploy/task.md)、[文档](docs/L2/LMDeploy)、[视频](https://www.bilibili.com/video/BV18aUHY3EEG/)| 100元算力点 |
+|第 4 关| InternVL 多模态模型部署微调实践 | [任务](docs/L2/InternVL/task.md)、[文档](docs/L2/InternVL)、[视频](https://www.bilibili.com/video/BV1nESCYWEnN)| 100元算力点 |
+|第 5 关| 茴香豆：企业级知识库问答工具 | [任务](docs/L2/huixiangdou/task.md)、[文档](docs/L2/huixiangdou/readme.md)、[视频](https://www.bilibili.com/video/BV1C5W2epEYP/)| 100元算力点 |
+|第 6 关| MindSearch深度解析及实践 | [任务](docs/L2/MindSearch/task.md)、[文档](docs/L2/MindSearch/readme.md) | 100元算力点 |
+
+### 1.4. 彩蛋岛
+
+||关卡名称|资料|
+|:-----|:----|:-----|
+|第 1 关| InternLM+WasmEdge 部署智能体实践|[文档]、[视频](https://www.bilibili.com/video/BV1K2rAYeEXG/)|
+
+## 2. 证书
+
+完成进阶岛闯关将获得精美的结营证书～
+
+![image](https://github.com/user-attachments/assets/86f420b1-5f82-4ae3-b7f6-4b4bdcdca1b8)
+
+
+## 3. 书生共学计划
+
+
+在大模型技术的浪潮中，面对混杂的众多信息，如何获取有效、可信的学习资源成为了一项挑战。  
+为此，我们推出“书生共学计划”，鼓励大家将实战营活动分享给你身边有需要的小伙伴，让每一位热爱技术的朋友都能在这个复杂的信息环境中找到自己的航向，帮助他们在大模型的学习之路上少走弯路。  
+
+
+参与方法  
+- 启航准备：通过填写问卷报名加入实战营，开启你的学习之旅。  
+- 专属海报：访问书生共学计划活动页面（https://colearn.intern-ai.org.cn/co ），系统将自动生成你的专属海报
+- 招募同行者：将海报分享给你身边的小伙伴，邀请他们报名实战营，共享知识的力量。  
+
+
+<div align="center">
+
+  <img width="440" alt="image" src="https://github.com/user-attachments/assets/99f03ded-da19-42a1-a42d-982fe58ae0c3">
+
+</div>
+
+
+
+
+独家奖励等你拿
+- 每邀请 1 位同学填写报名问卷即可获得 50 算力点。
+- 成功邀请 3 人，解锁 InternStudio 平台 24GB A100 及 80GB 存储使用权限。
+- 成功邀请 6 人，解锁 InternStudio 平台 40GB A100 及 120GB 存储使用权限。
+- 成功邀请 16 人，解锁 InternStudio 平台 80GB A100 及 200GB 存储使用权限。
+
+展现你的影响力，成为知识的使者
+这不仅是一个促进个人学习和成长的机遇，更是一个帮助他人、为自己赢得认可和资源的舞台。通过你的分享，我们可以一起帮助更多的人接触和了解前沿技术，期待你的加入。
+

configs/internlm2_5_chat_7b_qlora_alpaca_e3_copy.py

@@ -0,0 +1,219 @@
+# 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
+
+#######################################################################
+#                          PART 1  Settings                           #
+#######################################################################
+# Model
+pretrained_model_name_or_path = '/root/finetune/models/internlm2_5-7b-chat'
+use_varlen_attn = False
+
+# Data
+alpaca_en_path = '/root/finetune/data/assistant_Tuner_change.jsonl'
+prompt_template = PROMPT_TEMPLATE.internlm2_chat
+max_length = 2048
+pack_to_max_length = True
+
+# parallel
+sequence_parallel_size = 1
+
+# Scheduler & Optimizer
+batch_size = 1  # per_device
+accumulative_counts = 1
+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 = [
+    '请介绍一下你自己', 'Please introduce yourself'
+]
+
+#######################################################################
+#                      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,
+        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=load_dataset, path='json', data_files=dict(train=alpaca_en_path)),
+    tokenizer=tokenizer,
+    max_length=max_length,
+    dataset_map_fn=None,
+    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)