|
|
c9bb50d503
|
fix: train maskgit
|
2026-03-11 16:43:00 +08:00 |
|
|
|
c000b90794
|
feat: 训练时的 heatmap
|
2026-03-11 16:33:15 +08:00 |
|
|
|
22a2db464f
|
feat: maskGIT 加入热力图条件限定
|
2026-03-11 16:27:27 +08:00 |
|
|
|
326e6abf0b
|
feat: 热力图生成
|
2026-03-11 16:05:21 +08:00 |
|
|
|
0cb22e9cf7
|
feat: 分块掩码方式
|
2026-03-11 14:23:46 +08:00 |
|
|
|
8672c52ff5
|
chore: 取消注释
|
2026-03-11 12:30:44 +08:00 |
|
|
|
954c0109b9
|
fix: mask
|
2026-03-11 12:29:33 +08:00 |
|
|
|
fa72e18c1c
|
fix: maskGIT
|
2026-03-10 23:26:36 +08:00 |
|
|
|
d312fb65b7
|
feat: maskGIT
|
2026-03-10 23:06:23 +08:00 |
|
|
|
a119cbb155
|
fix: 报错
|
2026-03-10 18:24:40 +08:00 |
|
|
|
973434553a
|
feat: 为 transformer 加入起始与结束 token
|
2026-03-10 18:12:42 +08:00 |
|
|
|
d2ecc9e8c8
|
Merge branch 'master' of github.com:unanmed/ginka-generator
|
2026-03-08 16:06:14 +08:00 |
|
|
|
e66919d11e
|
chore: 微调 rnn-vae
|
2026-03-08 16:05:07 +08:00 |
|
|
|
22c98d54d1
|
feat: transformer
|
2026-03-08 16:03:55 +08:00 |
|
|
|
c8809b8ee7
|
chore: 调整调度
|
2026-02-13 00:05:38 +08:00 |
|
|
|
becf625bdb
|
chore: 调整部分参数
|
2026-02-12 23:50:08 +08:00 |
|
|
|
a07d2cf960
|
chore: 调整调度方式
|
2026-02-12 23:41:45 +08:00 |
|
|
|
1352d64a50
|
feat: 自定义调度器
|
2026-02-09 14:59:09 +08:00 |
|
|
|
45cfa3b611
|
chore: 调整kl参数
|
2026-02-06 23:54:59 +08:00 |
|
AncTe
|
dd9d8a3713
|
Merge pull request #1 from unanmed/feat-add-mps
use mps if available
|
2026-02-06 23:50:25 +08:00 |
|
tocque
|
2357363d76
|
use mps if available
|
2026-02-06 23:21:59 +08:00 |
|
|
|
cf1cbbc680
|
fix: 验证报错
|
2026-02-06 16:51:18 +08:00 |
|
|
|
dc6d1c69be
|
fix: 验证报错
|
2026-02-06 14:35:31 +08:00 |
|
|
|
948132797d
|
refactor: 换回 vae
|
2026-02-06 14:31:07 +08:00 |
|
|
|
9887abcd01
|
fix: zero_grad
|
2026-02-06 01:23:58 +08:00 |
|
|
|
b01357d99e
|
feat: 更改调度
|
2026-02-06 01:20:27 +08:00 |
|
|
|
bc73cf9cc3
|
refactor: vae 退化为 ae
|
2026-02-03 13:35:52 +08:00 |
|
|
|
df2bd80bec
|
chore: 调整数据集
|
2026-02-03 13:26:34 +08:00 |
|
|
|
10166fa073
|
chore: 调整参数
|
2026-02-03 13:25:41 +08:00 |
|
|
|
f22afb5f72
|
chore: 调整超参数
|
2026-01-20 23:57:05 +08:00 |
|
|
|
05b3b7c171
|
feat: 编码器特征融合添加归一化
|
2026-01-20 23:53:31 +08:00 |
|
|
|
b7fe24ee4c
|
feat: 编码器加入 transformer 整合，改进解码器
|
2026-01-20 23:48:25 +08:00 |
|
|
|
8caa37a144
|
fix: 验证报错
|
2026-01-20 19:05:00 +08:00 |
|
|
|
4d244d021a
|
feat: 调整 decoder 的位置编码
|
2026-01-20 16:50:11 +08:00 |
|
|
|
14f391f4f4
|
fix: vae参数错误
|
2026-01-20 16:43:59 +08:00 |
|
|
|
dd6a043487
|
feat: vae编码器换为 rnn
|
2026-01-20 16:31:13 +08:00 |
|
|
|
169a514dd1
|
fix: 插值图片显示
|
2026-01-20 00:00:24 +08:00 |
|
|
|
f22f06ef2d
|
fix: 插值文件名
|
2026-01-19 23:54:11 +08:00 |
|
|
|
83d6c31704
|
chore: 调整 batch_size
|
2026-01-19 22:34:08 +08:00 |
|
|
|
79cf3ab226
|
feat: vae 模型
|
2026-01-19 22:30:00 +08:00 |
|
|
|
62ec7daa16
|
fix: typo
|
2026-01-19 00:46:13 +08:00 |
|
|
|
cfc022724a
|
feat: rnn 随机采用自身生成的内容
|
2026-01-16 23:01:37 +08:00 |
|
|
|
d0decfc63a
|
fix: device 不正确
|
2025-12-17 13:12:44 +08:00 |
|
|
|
bf5160edac
|
feat: 添加地图蒙版
|
2025-12-17 13:09:10 +08:00 |
|
|
|
1962c7a712
|
fix: 损失值计算
|
2025-12-15 12:41:39 +08:00 |
|
|
|
3946d83d6c
|
fix: 交叉熵损失计算
|
2025-12-15 12:32:14 +08:00 |
|
|
|
102e19cefb
|
chore: 添加注释
|
2025-12-15 12:19:57 +08:00 |
|
|
|
bd821aa8e4
|
chore: 添加 dropout，调整调度与损失值计算
|
2025-12-15 12:18:34 +08:00 |
|
|
|
616b7fd39c
|
chore: 微调模型
|
2025-12-14 12:06:53 +08:00 |
|
|
|
7ae4ae4572
|
chore: 增大 RNN 参数量
|
2025-12-13 21:34:59 +08:00 |
|
|
|
4d1179bcfb
|
feat: 训练时使用 target_map 做局部提取而不是自己输出的内容
|
2025-12-13 19:57:06 +08:00 |
|
|
|
d42c2eee43
|
fix: 修改超参数
|
2025-12-13 19:46:09 +08:00 |
|
|
|
ae53194694
|
chore: 修改 batch_size
|
2025-12-13 19:14:48 +08:00 |
|
|
|
27b8c56cd2
|
fix: inplace
|
2025-12-13 17:48:51 +08:00 |
|
|
|
771553b8b8
|
feat: rnn train
|
2025-12-13 17:38:57 +08:00 |
|
|
|
fa8ded2ecd
|
feat: rnn 训练
|
2025-12-13 13:13:22 +08:00 |
|
|
|
c79662089b
|
feat: RNN 融合编码
|
2025-12-13 13:03:19 +08:00 |
|
|
|
1ccac9e60d
|
fix: rnn
|
2025-12-12 16:41:27 +08:00 |
|
|
|
7fb1625e0b
|
feat: rnn
|
2025-12-12 14:57:06 +08:00 |
|
|
|
310b3fae80
|
feat: 适配新的数据集
|
2025-12-11 22:55:55 +08:00 |
|
|
|
224005b44b
|
feat: 自动标注过滤
|
2025-12-10 19:54:10 +08:00 |
|
|
|
5586ea1039
|
chore: 微调模型
|
2025-06-14 15:06:09 +08:00 |
|
|
|
fb0323d874
|
perf: 改进生成流程与损失计算
|
2025-05-15 17:44:29 +08:00 |
|
|
|
fa48863946
|
perf: 模型微调
|
2025-05-11 23:50:08 +08:00 |
|
|
|
21b693ec21
|
feat: 图卷积部分换用 TransformerConv
|
2025-05-09 13:07:03 +08:00 |
|
|
|
7b138c66d9
|
perf: 加强 GCN 部分
|
2025-05-08 18:42:15 +08:00 |
|
|
|
55f09fb37b
|
perf: 改进输入部分
|
2025-05-07 15:38:31 +08:00 |
|
|
|
d800a2382b
|
feat: 当前阶段条件注入
|
2025-05-02 15:20:58 +08:00 |
|
|
|
e3e496957c
|
feat: 损失值改进
|
2025-05-02 13:59:56 +08:00 |
|
|
|
53041ab754
|
perf: 改进网络结构
|
2025-05-01 22:08:39 +08:00 |
|
|
|
10e6ad6394
|
fix: 训练循环 & 怪物数据处理
|
2025-04-30 21:35:53 +08:00 |
|
|
|
dc1eb8fa92
|
feat: 提高标量数量
|
2025-04-30 21:00:45 +08:00 |
|
|
|
44d6c7410e
|
feat: 数据集密度占比
|
2025-04-29 21:57:55 +08:00 |
|
|
|
8add448195
|
feat: ginka 数据集生成工具生成新版数据集
|
2025-04-29 20:47:05 +08:00 |
|
|
|
44b90e7630
|
refactor: 修改目录结构 & feat: 条件注入
|
2025-04-29 18:24:01 +08:00 |
|
|
|
a28e56456d
|
Merge branch 'master' of github.com:unanmed/ginka-generator
|
2025-04-28 22:52:00 +08:00 |
|
|
|
afde7be592
|
feat: 新的数据处理方式
|
2025-04-28 22:51:53 +08:00 |
|
|
|
30bf95d493
|
feat: save model as deployment
|
2025-04-28 14:26:54 +08:00 |
|
|
|
5be9decf95
|
perf: 优化训练流程，改进判别器与损失值
|
2025-04-27 14:39:03 +08:00 |
|
|
|
a94b07bda8
|
feat: 改进训练流程
|
2025-04-20 22:06:20 +08:00 |
|
|
|
87016c67e8
|
feat: 改进输出头和特征融合
|
2025-04-19 15:05:26 +08:00 |
|
|
|
ef0b7ffba2
|
fix: 修复损失值计算
|
2025-04-16 21:54:53 +08:00 |
|
|
|
9b5abf177c
|
feat: 添加 js 散度损失
|
2025-04-16 18:39:17 +08:00 |
|
|
|
447c28ff5e
|
perf: 修改损失值计算方式
|
2025-04-14 14:15:13 +08:00 |
|
|
|
f6b1ad6ebd
|
refactor: 课程训练
|
2025-04-13 21:06:07 +08:00 |
|
|
|
99f46150be
|
perf: 改进 WGAN 训练
|
2025-04-10 22:42:58 +08:00 |
|
|
|
268b21e0b7
|
perf: 调优超参数和模型
|
2025-04-07 16:41:27 +08:00 |
|
|
|
8130296e1f
|
perf: 调优部分超参数
|
2025-04-06 21:19:18 +08:00 |
|
|
|
29cfb4d029
|
refactor: 改为 Wasserstein GAN
|
2025-04-06 18:44:18 +08:00 |
|
|
|
d7209a68a2
|
feat: 记录训练过程
|
2025-04-03 13:36:05 +08:00 |
|
|
|
6a1aeaa77e
|
fix: 浮点数被意外转成 LongTensor
|
2025-04-03 12:12:13 +08:00 |
|
|
|
325eb599c3
|
feat: 支持流式传输，改进数据集
|
2025-04-02 17:57:02 +08:00 |
|
|
|
8dea79a9f0
|
feat: 使用 IPC 通讯并实现单脚本对抗训练
|
2025-04-01 22:41:23 +08:00 |
|
|
|
4721e9a141
|
feat: 改进生成器网络
|
2025-04-01 12:53:52 +08:00 |
|
|
|
96f828e29b
|
feat: 修改 ginka 模型的损失值计算
|
2025-03-31 13:46:33 +08:00 |
|
|
|
4f7dbb6fb3
|
fix: 训练报错
|
2025-03-30 17:01:59 +08:00 |
|
|
|
5669f49af0
|
refactor: UNet 部分重写并改为条件注入模式
|
2025-03-30 13:48:01 +08:00 |
|
|
|
49ee543732
|
fix: 数据集报错
|
2025-03-26 22:44:11 +08:00 |
|
|
|
8f892fc7f4
|
feat: 优化图块数量损失函数
|
2025-03-26 21:25:09 +08:00 |
|
|
|
a5510ef211
|
feat: 改进模型与相似度计算
|
2025-03-26 21:06:28 +08:00 |
|
