KMnO4-zx
|
30f3f01619
|
refactor(dataset): 使用tokenizer动态生成a_sequence并替换硬编码值
fix(ddp_sft_full): 修正参数默认值和优化器类型
docs(ddp_pretrain): 添加详细注释和优化参数描述
|
2025-06-21 11:39:40 +08:00 |
|
MengYue-MK2000
|
b1ac936d36
|
created windows_download_dataset.sh, deleted original changes in download_dataset.sh
|
2025-06-19 17:52:24 +08:00 |
|
Reagan Zhang
|
18ff1a73a8
|
Update download_dataset.sh
Update Mac installation for modelscope
|
2025-06-19 16:09:59 +08:00 |
|
Reagan Zhang
|
56fb0c34d4
|
Update download_dataset.sh
|
2025-06-19 16:06:05 +08:00 |
|
KMnO4-zx
|
ce535629ca
|
docs(chapter5): 更新模型文档并添加数据处理脚本
- 更新LLaMA2模型文档,修正图片引用和编号
- 添加Attention结构示意图
- 新增数据处理脚本download_dataset.sh和deal_dataset.py
- 优化文档中的代码示例说明
|
2025-06-18 16:26:33 +08:00 |
|
KMnO4-zx
|
ada2e0c44f
|
fix(download.py): 修复解压命令未指定目标目录的问题
|
2025-06-18 12:34:52 +08:00 |
|
KMnO4-zx
|
9efbb69dfd
|
docs(chapter5): 添加LLaMA2结构图并更新依赖
更新requirements.txt中的pytorch为torch以保持一致性
|
2025-06-09 22:14:01 +08:00 |
|
KMnO4-zx
|
9569c9fdca
|
fix(tokenizer): 将add_prefix_space配置改为false
|
2025-06-08 09:27:21 +08:00 |
|
KMnO4-zx
|
32c3f16b8c
|
fix: add chapter5 reauirements
|
2025-06-03 18:42:51 +08:00 |
|
KMnO4-zx
|
3512f55993
|
update ch05
|
2025-02-26 20:31:51 +08:00 |
|
KMnO4-zx
|
ca3e727e1c
|
update ch05
|
2025-02-26 11:24:19 +08:00 |
|
KMnO4-zx
|
9e6d8a3f77
|
Add: ch5.3 code
|
2024-09-22 16:02:14 +08:00 |
|