diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000004.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000004.png
new file mode 100644
index 0000000000000000000000000000000000000000..8c5d570cba17be6ef960ad2942ed5802c805a80e
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000004.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16efa95286028792e4fa4e73470f77b38193d37048f2ab321b799993fda96e34
+size 2252877
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000013.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000013.png
new file mode 100644
index 0000000000000000000000000000000000000000..d2507d2c7380d2e94b606ff49299d16ffb96f626
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000013.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b991836db3dc841f679adb400c35868495bce66b4172835ec41d0fbe79ad3ba9
+size 1379007
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000034.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000034.png
new file mode 100644
index 0000000000000000000000000000000000000000..f487144c6cfad3fb81377633bbb427ed7fcbaa63
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000034.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf7fbde9f5be6afbb3842ba6d73902cb9b52e28da6698c57cfa25208c72b707b
+size 1739300
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000038.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000038.png
new file mode 100644
index 0000000000000000000000000000000000000000..ae7d333c9860176705d166021823c498a9f4cfc0
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000038.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6727be5418afb662f99a853e51b2f4ac6d7c28464b7e4c0d99c4bc987ad5a559
+size 1579560
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000040.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000040.png
new file mode 100644
index 0000000000000000000000000000000000000000..76873515fbc6aa564df29354ad8d006334737d23
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000040.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b28a5b57981f17196bf9f103ef95f8f5fbe9f018c6b82c2cdd716fa98a64f2b4
+size 2019844
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000055.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000055.png
new file mode 100644
index 0000000000000000000000000000000000000000..a7540dc62468dcc2e38fc7608c7a3fe6c5c1d5de
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000055.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f99c06351d5c2024da81c677fa6a41cd0887c5b755c9918a400414157e923460
+size 1702399
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000061.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000061.png
new file mode 100644
index 0000000000000000000000000000000000000000..bc4b0b158a3f73ae4cabfa8ba92d1e51be48270c
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000061.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d92e4075f42ae1cb7baed71482a70b64e5fd900380741fbe3836237a2a5b70f
+size 1674328
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000089.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000089.png
new file mode 100644
index 0000000000000000000000000000000000000000..560c4722a23c988f400a2d7771c4195fa820f36d
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000089.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2de297fa9241e07677794cccf6d8be116b26a0ef333106d4bb0cf90d9eab99a
+size 1716085
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000096.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000096.png
new file mode 100644
index 0000000000000000000000000000000000000000..70f4b3a8ffe2cef414fcf3d20de826c056b6a141
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000096.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc9c16d443a61c5384f00109341c3880af596d0d21156ecb78d00203f56c865a
+size 1693671
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000110.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000110.png
new file mode 100644
index 0000000000000000000000000000000000000000..1a781e8b204144ef7e5ae83cf1f84cc91b99b5ee
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000110.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f435a741db794f9c1389c9b6203e4c16fbf37cd142b5c000e4fa3594a0c3dfa
+size 1965984
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000114.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000114.png
new file mode 100644
index 0000000000000000000000000000000000000000..17f9e88451bdeb5be3d00f1694fbd1a06ed796cf
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000114.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28b1ca5b23e1d9b4391c1a9afaf347ed6c873cb08a5c00d10c043d3185a0aade
+size 2027343
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000131.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000131.png
new file mode 100644
index 0000000000000000000000000000000000000000..86fa2ba06f569bcd1c6b7b89fef773c0cf9c95db
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000131.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77d9c591080bfc34dc3688dbfbcdc4b4bd838b667d674fea0eb43423950bcf0e
+size 1807625
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000136.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000136.png
new file mode 100644
index 0000000000000000000000000000000000000000..81b7aa2240d553e56dbf8d96e64aa4b47fa807e7
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000136.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a2360a91959b2933a7218f944168ce8a069159fdd38415399ef3c58d592fcc53
+size 1562566
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000146.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000146.png
new file mode 100644
index 0000000000000000000000000000000000000000..146c2d17e0f9fbc83ddf3627aa981034c2d0747c
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000146.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4aef6821f175ffffb714881935979d008b25b6d396cc9f61a2e49fe8ef2184b1
+size 1668778
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000155.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000155.png
new file mode 100644
index 0000000000000000000000000000000000000000..cec359af83ccb4e8505f0160b2f956dc6d3beaed
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000155.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:acb12fe3aff37e3c33b115d09cbcff4e29f515e813872370bd3a6b8af8e37cbc
+size 2210330
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000157.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000157.png
new file mode 100644
index 0000000000000000000000000000000000000000..3d2db63f065982ffbea742c04bcbd85722d75963
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000157.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0454bab487a41107e4a9e2f52aa878bc5e4e673390ce0e35267dd457f6c2201
+size 2301962
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000163.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000163.png
new file mode 100644
index 0000000000000000000000000000000000000000..401835ce8d7a65078419e90f447a67b19be280e7
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000163.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab60c6e5f6f74330b5e3f0338f3ba1616f22e280b285dd9a17636b5b590b4302
+size 1614893
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000181.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000181.png
new file mode 100644
index 0000000000000000000000000000000000000000..d3124bdd56d4197aaf7c11a8db7494a9b9335b03
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000181.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0065a16ac15d774d7bbf21ba79e21dec3bc49a2823de8f16c8aac42e90a5d60a
+size 1906419
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000187.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000187.png
new file mode 100644
index 0000000000000000000000000000000000000000..7608698faf42efd0c97582e5d234ad07b649e9e2
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000187.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be908473ad828d843fb008d9697bda2092873f36b1f8ef31b7491ac8543031ce
+size 1967767
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000192.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000192.png
new file mode 100644
index 0000000000000000000000000000000000000000..85a64fd7c8b240a532bad6c221cad64b22fcd5c7
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000192.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0496d2c506ab9911fe8f276937442f99e4cc5c96aaf3873c1f7615fe80abd7f4
+size 1365582
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000209.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000209.png
new file mode 100644
index 0000000000000000000000000000000000000000..03844a9f9f3337d7cc0717997e28630c63f60a57
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000209.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c16729b63ac39e2d912fed3fd2a26349fe625d5a52b2ff1f8258c3eb1bf4895
+size 1072240
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000232.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000232.png
new file mode 100644
index 0000000000000000000000000000000000000000..218e5fbf9cf3b4ddc3457f43f0b33930dabd8bcf
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000232.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ade942339f4316f238852230f59b0742b23e923852d6eca1561f3f61a4712942
+size 1703202
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000241.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000241.png
new file mode 100644
index 0000000000000000000000000000000000000000..2f3f0130f512d62aa73879f3e73581c4c86245a1
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000241.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2789aff4acf4ac924f6166783b846a7ef1738f98ac1f4625ec5e3aea8771bbdc
+size 2401931
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000285.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000285.png
new file mode 100644
index 0000000000000000000000000000000000000000..1dbdcddfc0bba067f611a09534888a7070b8be34
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000285.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2773ce588fffe577b30689d8b3b12badbc7cb6e305ff9ca49b2364d6c5db5c3
+size 1792340
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000301.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000301.png
new file mode 100644
index 0000000000000000000000000000000000000000..8146e04c0acb7cb2d43c1ededb2172f3eb79ffb2
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000301.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87d581fffbcf6eba5e8066bd6fa6d31b6a4185ac181f815383556b547a54ea66
+size 1776206
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000330.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000330.png
new file mode 100644
index 0000000000000000000000000000000000000000..64976bd66622b710859c6ede8d411f5782100b6d
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000330.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:178b768f0f2d58e0698490dadcc7d089d316a8270a26cc6f2988375d97b36b0d
+size 1478023
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000379.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000379.png
new file mode 100644
index 0000000000000000000000000000000000000000..c82a69a1125077bdc8e075c93abc47612bad9e99
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000379.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2208f19a06a01f73c928d38ea89cd01c099b26761f39167d1cd8f77bda274986
+size 1551891
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000383.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000383.png
new file mode 100644
index 0000000000000000000000000000000000000000..531d20613190cca5e2f41897565a3ec4751efc2a
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000383.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a25e31081a1e5535d8300696150700c42b3ddf3e6593e3a1dd67a6b77174d1a
+size 1532620
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000389.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000389.png
new file mode 100644
index 0000000000000000000000000000000000000000..a9f9fc72d9953ed5e7cf5e191cae4b22d52014e6
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000389.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce0b11c1950a310e7f9c838372dbc086a7e5554a32c474d998a8abaa3514f790
+size 2004654
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000407.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000407.png
new file mode 100644
index 0000000000000000000000000000000000000000..b7d5765c2c7e442e7ed360fda9f535dd0674f956
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000407.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c783b0a8b2a77359a759c466b1d117b7fd6d151a610535708c0d76a0f031ad6d
+size 1493071
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000415.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000415.png
new file mode 100644
index 0000000000000000000000000000000000000000..a62aeaa022d79c033f8a79298179ea46b4148152
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000415.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a0f02ff990c5aa94b1815c52baca72a235635a9617cf08d9e587c7d69534993
+size 1447742
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000454.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000454.png
new file mode 100644
index 0000000000000000000000000000000000000000..bdf413ea455603655e511c1382dcf178f45c5c6e
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000454.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af2f5b09dff4a9a608b33b3f1684dde8a4d82a902d5fa8b12db3d2342240c797
+size 1312345
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000455.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000455.png
new file mode 100644
index 0000000000000000000000000000000000000000..24ef0148a5b9124f34258e63e625db3dc9336ee0
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000455.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b3ee009ac53535182a1d33f1421374bb274c11488756625ae51f92729ba97da
+size 1914097
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000466.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000466.png
new file mode 100644
index 0000000000000000000000000000000000000000..de3e1c0c038cc4d1a48750169bc57d063b5873b2
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000466.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1306c4f6ddab57a70cd3e36d7462d3e41a315db81f21410c0de83045178d336e
+size 1846582
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000473.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000473.png
new file mode 100644
index 0000000000000000000000000000000000000000..bab037d4b51bcb32bc0bd8d682d508c82d9e91b9
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000473.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c0881ef81a53afc63fab7e3bd412dccb578547f10b830cc094d3dbce1d7422b
+size 1778974
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000477.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000477.png
new file mode 100644
index 0000000000000000000000000000000000000000..f000e84ab4fa0faa41d40572904e05669c4e3050
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000477.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c512a9945f946d91c2c6bef666fd1d4985b36b199ff9ddc5fe947b0fceef2ec3
+size 1334559
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000485.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000485.png
new file mode 100644
index 0000000000000000000000000000000000000000..645e8d8b8ae79c958fc57497acc4f190cf5ac58a
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000485.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9788ae5689ef7806f3583b04293bd36ba0e440eaeffd4724bad9b5d7b6baa874
+size 2048242
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000486.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000486.png
new file mode 100644
index 0000000000000000000000000000000000000000..1978dc02ed7a0a6d5eb02b308e454977618d79fc
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000486.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c378a912f91a62715428b13d37e3ce2afefe5528a6437b117a3948aff6221307
+size 2277765
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000490.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000490.png
new file mode 100644
index 0000000000000000000000000000000000000000..750d6a0ba01c15375364079f2eb8f64441c4f041
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000490.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35f8d03ec9b6caf0d217a208417a7b24f56449bb6c9267ab260e178bab08b55c
+size 1949544
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000501.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000501.png
new file mode 100644
index 0000000000000000000000000000000000000000..bd77cc62f07acf5fca2ab6b10d52f6e3df0d0439
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000501.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e61fe5b9ef3547f3f7b766e499760e738bc555693be1015fa5160f31c9e212ce
+size 1450874
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000532.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000532.png
new file mode 100644
index 0000000000000000000000000000000000000000..70e4ff827e7af18f25b293ee29bf1ac61d1e3eac
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000532.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd7ce6b37bfda974005f77bc35844b6c45f8d3f9b6d58a85ca3c12a80bdac262
+size 1810493
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000540.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000540.png
new file mode 100644
index 0000000000000000000000000000000000000000..a98c76788e7ec280ffc6f2bfa61545e01bd199e6
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000540.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f796b5db526d3d0ea28975d043f24769bd85aec9d4c7b2a930dbd5e181ac8b1
+size 1733793
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000547.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000547.png
new file mode 100644
index 0000000000000000000000000000000000000000..adfb1c82533e80a326799359b35b4868b8d52130
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000547.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1558d126fcef641d3f526a78e69cfb478a77e76e2ebc317b6a8c31717f71d076
+size 1774972
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000559.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000559.png
new file mode 100644
index 0000000000000000000000000000000000000000..e2035ced3af34f0802c4f8df93ffa7cc3be8aa4e
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000559.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32ccc06634dba42f70aa4a8775c4f3c455650ef8ce19715e7cd72b5ef3655f2f
+size 1807654
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000563.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000563.png
new file mode 100644
index 0000000000000000000000000000000000000000..d9263e5852693f1d807907ed056a47d968485cfa
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000563.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a9448472c50c646ac288aec7e42e8fd5e1ff28f8cc5d8a6c0fa2e1eb476fa7a
+size 1800284
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000611.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000611.png
new file mode 100644
index 0000000000000000000000000000000000000000..edb44187c67f85b82eae8721d877c4a45b89c7f3
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000611.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e7f001fb11bb2409bece0e2a1f37f497fcfed4f1ef2c44355dcf5233821376e0
+size 2334955
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000639.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000639.png
new file mode 100644
index 0000000000000000000000000000000000000000..45137b8e1d5f4878c9c748a3d623c7c51ea7dcc8
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000639.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a006e24e1579c5ac8b88e305bd365fa4af609c70c64c6bbde2c4ca942935829c
+size 1284498
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000653.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000653.png
new file mode 100644
index 0000000000000000000000000000000000000000..e8912523f04c8f1c88ae4df1d375dd8dcca94202
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000653.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f30e6ddde6ab38f8f78d3700c6e761e954d040f88be71ef8beec91f7e99da546
+size 2432087
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000657.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000657.png
new file mode 100644
index 0000000000000000000000000000000000000000..dfec002850d1cb0fe8109e4b643c4e904f389e4e
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000657.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:217a36e76c026fc3c6929451f541ce000e59094217b23e3a81ebefff7c8ed1c7
+size 1334674
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000658.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000658.png
new file mode 100644
index 0000000000000000000000000000000000000000..13e011ee76b8a313fc2ce7e863eab0281ff6c3ac
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000658.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b05f0e4209d42807375008b0fdc47823e6e0519e0d3bb39a94ab4ad00ded2e68
+size 1951775
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000669.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000669.png
new file mode 100644
index 0000000000000000000000000000000000000000..9f4df2b01f10bf73ed0c8762231c5ee64aaafe3f
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000669.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c7872bb399dd02b2c78d0d03b622f869f1c3e5d819b5d1bd08a7f70d26d64b5
+size 1359468
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000673.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000673.png
new file mode 100644
index 0000000000000000000000000000000000000000..99e0efe62530eaf6357aae72aef60eae72783538
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000673.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7653cda9c7e327afead6998ea250866ded31a5b8c7f8728f6a99bd85be365c7
+size 1474896
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000681.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000681.png
new file mode 100644
index 0000000000000000000000000000000000000000..500440024220bb1cf25a2018ce0279c488915f3a
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000681.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7fee6c973d2b04374a5c925b131e3a9db6c3ffeb98ad94d4f9145c866c34f915
+size 2408275
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000690.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000690.png
new file mode 100644
index 0000000000000000000000000000000000000000..72b7a6d0ae9460031650fed29dd37de38be1a63c
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000690.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23cf6d86b87d8b49b7d358a1198da3d3c9a67cbd0e71c175e61e4d2122cfe0c6
+size 2454118
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000707.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000707.png
new file mode 100644
index 0000000000000000000000000000000000000000..4ff3c08d6adbd118faf38cb3074bd81f41bc9800
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000707.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3d35e6a1aee78b732013050328943af3aaa6fcf4e2ba9aada3884f2f6e32f96
+size 2578520
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000732.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000732.png
new file mode 100644
index 0000000000000000000000000000000000000000..dc4eb82c2c7156ea7b7116ee8d644df047935096
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000732.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93f389790059730c8fd86f84e3999963501dc8928e08816ccae951fa149c7161
+size 1685580
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000748.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000748.png
new file mode 100644
index 0000000000000000000000000000000000000000..79fdb9106aecc5cad9e23ead5a1a1e85ab1b438d
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000748.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96033abcb3565d12b4c3a7c4dd6b7b645c9c7662e79bab8b4ada99634353cb94
+size 2198196
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000761.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000761.png
new file mode 100644
index 0000000000000000000000000000000000000000..70106225f749f5bee174a2e9bd9f4084ac05b2d5
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000761.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6fa3e4bd96f3317c80dad1b9d1854fc6d644fea9ab5331ce980693aa1be1c16
+size 1590288
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000770.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000770.png
new file mode 100644
index 0000000000000000000000000000000000000000..d048fa678a4b11f7cd222b6fccc822cc1226618a
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000770.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac8973c81146e54b7a88bd8e319b4b7f40e3f5cd6dea26d95fb6744b8584fcb6
+size 1587163
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000797.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000797.png
new file mode 100644
index 0000000000000000000000000000000000000000..9eba4b082cd03dc4cb29d4103c59440325db1443
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000797.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:90756a66d0f7185ce0143a4a172148c1bcfa73e3ff619971662503ba05ef58e6
+size 1952409
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000813.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000813.png
new file mode 100644
index 0000000000000000000000000000000000000000..bbb8649aabcb02399a9d96fef88345529a3d65da
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000813.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b9deb8ae4a52a7cace33fd6bea7628a1d3e2a0ea778de393efe2a0f8e32193e
+size 2088728
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000816.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000816.png
new file mode 100644
index 0000000000000000000000000000000000000000..839d938db837081e5033f27e97e85f1fcf91b971
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000816.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4db40d6591ab1d7f5d121ee776ddd65f773af1281150c97a6748a21ef4b955c0
+size 1866032
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000838.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000838.png
new file mode 100644
index 0000000000000000000000000000000000000000..39d85b6937735f479fcdcf002afc6cf2461bf695
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000838.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a6c7c2d213ff45b14f9cd3f6bc72ecf107740a8ec9a11781a57e7e07addcd0b
+size 1632500
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000856.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000856.png
new file mode 100644
index 0000000000000000000000000000000000000000..707545c2c243593ea3bbc87c72d01861777e327e
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000856.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:709c9c3e863d21fd7c740db07b62a3ec00c0b75267ef47f5555e116fcd8df6c5
+size 2384868
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000867.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000867.png
new file mode 100644
index 0000000000000000000000000000000000000000..c7c0c8a40ca68b36c8efcf293300127eddaf3017
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000867.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:333f09ec3c6ab55dce09adbb1e6ee6621d03b5d83c9075cb7d53d986933cfea1
+size 1729388
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000888.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000888.png
new file mode 100644
index 0000000000000000000000000000000000000000..4e32ef102f585ae9e3a8a97964ff7be204fd305f
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000888.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62952c2d4ac607173b1ec66a0b2c7da4bf8e6dbba7729e63fefeca72ed54603a
+size 2566730
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000890.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000890.png
new file mode 100644
index 0000000000000000000000000000000000000000..b5e9ae8563e9b6d491b5ddc910a373a36f055aca
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000890.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e1c5b89070387a5b7e89ff51e77df343b3d240051b9b7d51a9f3677df13a9f5
+size 2327842
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000902.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000902.png
new file mode 100644
index 0000000000000000000000000000000000000000..9006ccb315cc8247a2ce0092c35cdd49f06b938e
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000902.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88407bf40d4fdaaf3d772d3bac08820378be5ef92ef8beee171ad54ced18bb5e
+size 1851305
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000925.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000925.png
new file mode 100644
index 0000000000000000000000000000000000000000..2b51defbef5ead5f84a7aae2e8cd8b552667afa4
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000925.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce3f6099a7d06137df901df30583ac4b90919252dfb9efb0945dcb3b4cfa7643
+size 1392320
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000935.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000935.png
new file mode 100644
index 0000000000000000000000000000000000000000..8d42e5b6de178f78afbfccaa601e584b8a863f8a
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000935.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1dfc9068d084509dff912874afb13d0b57931f5f7608856416a09ae86ac1b67e
+size 1473494
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000941.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000941.png
new file mode 100644
index 0000000000000000000000000000000000000000..db00578631f464fb7444640852faf7edba45bf20
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000941.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1c1c79ccaf7a53464a77c2d5eabfcbdf0954e108b0468bb98fd7fa0d9fc8081
+size 1278955
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000945.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000945.png
new file mode 100644
index 0000000000000000000000000000000000000000..f6c588f3170e95ede28ae3952627f613891a0c0b
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000945.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5dc3264698116f3244f0279b867ed00520c778441a35f74dd119f010982bcef5
+size 1475460
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000959.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000959.png
new file mode 100644
index 0000000000000000000000000000000000000000..77374092ca8f47e16ea135375174fda7fd469496
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000959.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0b15e0ccddc2037575a4ad6e653dda5957a84a08bf2e49d8c2202db748593ea
+size 1813037
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000986.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000986.png
new file mode 100644
index 0000000000000000000000000000000000000000..4f385a249657363ab1a27484395ab30e94d855e6
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_000986.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c253b0626ffdb78f7223882095e4f16a32e3c0f4801465523c08e4a136410191
+size 2544616
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001003.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001003.png
new file mode 100644
index 0000000000000000000000000000000000000000..d06535e1e116fb1cba5d6a9a9d429277ee95c871
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001003.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3edaf583aa797c011f99396fca1f92e204c168625268f599028e83b87a0ffd2a
+size 996888
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001008.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001008.png
new file mode 100644
index 0000000000000000000000000000000000000000..3f0f84f5f78128948bb8e0bd769f12b6f5e4d956
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001008.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5bc8df3aba446de7ff09b4aca064ed12dc36d2c6bda492ba5d199a90405f7be
+size 1420164
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001013.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001013.png
new file mode 100644
index 0000000000000000000000000000000000000000..6cc0d10856def8292c10b6bd23c99692a9bb9677
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001013.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa22a9df50d1f568bc18e6fcd5dd3d4a8e7db07dc8f7f52425871defde05021d
+size 1559430
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001068.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001068.png
new file mode 100644
index 0000000000000000000000000000000000000000..d4b69107dcdfa73f1110d9a9e5266398cd47a67b
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001068.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:755dcf3d908902d9598d135a32e7d4801b5fb008af4884eba37bce3e138227ed
+size 1457647
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001069.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001069.png
new file mode 100644
index 0000000000000000000000000000000000000000..8419c67a08433b2c898349333f510dc1d26015f9
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001069.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ebe3bc0b3432da44dd3e2c8693c94bca2b0779185fddf52031c3540bc62f9f28
+size 1471271
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001095.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001095.png
new file mode 100644
index 0000000000000000000000000000000000000000..6ca97aef60545c92000b6cf0f16a068ecf9c66a7
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001095.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68fb926ed5b51094b73abed11cac2c68e688becb85b7529dd31439cbbc1d5f9b
+size 1538351
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001122.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001122.png
new file mode 100644
index 0000000000000000000000000000000000000000..f474167767eacb7c10aab08e8fd520c2172df2fe
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001122.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5828c03bd936c48b6cb54a9e3a2442b9da2fdb2a6750af5695e1a1f3914c53f
+size 2335617
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001123.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001123.png
new file mode 100644
index 0000000000000000000000000000000000000000..ff04ce98289e90a46d267e0a329bf1f1c1a38cd9
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001123.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:479079aec5631f64785294de5d3c7c23dc5d926e881fe5e679b93282df62c3f8
+size 2105757
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001133.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001133.png
new file mode 100644
index 0000000000000000000000000000000000000000..1165f19968bad7f0d1dbab23b7fd58c000f25ac0
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001133.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6bf9eb2d0efec1797bb601698f59f34aaab91858aecba8b5c7d7eafd4a25a9b7
+size 1279158
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001143.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001143.png
new file mode 100644
index 0000000000000000000000000000000000000000..e2ba311c0b537a1cd4fd942894b47c39d377e9e9
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001143.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e41d1632f30221bce34acbc3d6e56a9dbe7a6ea312ce5a60292b86620384087
+size 1614331
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001182.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001182.png
new file mode 100644
index 0000000000000000000000000000000000000000..70b3ed5085df663b70ec93aa798bfbd88655316f
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001182.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52150c11776aa9550a48bc43127b51761b9ded2099ac962b80fc4f8a14ace0d0
+size 1105294
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001195.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001195.png
new file mode 100644
index 0000000000000000000000000000000000000000..7570e3e972bad7cddba2e183fd47d471738bf177
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001195.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7bc22626e50310c9ade77a2ce760f9fd0add82141ee280a2a2563d0173db4e2
+size 1409959
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001204.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001204.png
new file mode 100644
index 0000000000000000000000000000000000000000..c70768503d8faaeae17d9a2a66814e9d5c1c2387
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001204.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16c1d38062665f22b28b93cb1f17d908eda217ea048e4d6dd20fd1c00d62cb67
+size 1280406
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001247.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001247.png
new file mode 100644
index 0000000000000000000000000000000000000000..60896771c1bc169a027b3d898d0abe5c883f9d79
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001247.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f7df7e61a760df96c3097f55f66fd99641bc46bcb5b8710e4fee7f1fa2b2cb0
+size 1540385
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001251.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001251.png
new file mode 100644
index 0000000000000000000000000000000000000000..30fcd692fd109c51b6c5395e501ea2e793611e0e
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001251.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6d4b0f910aefc0241229e077adf5a192292a67d253e178f29f30beee12fc2365
+size 1415186
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001278.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001278.png
new file mode 100644
index 0000000000000000000000000000000000000000..0b6f417d71cb52edd75ae075feeea153ff509cb3
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001278.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24ca26531b515cd22dc102923577b766d57b0b5045039c9f10a689e8bfd6f639
+size 1515604
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001302.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001302.png
new file mode 100644
index 0000000000000000000000000000000000000000..dfa7cb3c40b6973a2cdad650b5f02be7448a22d9
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001302.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d8866b14a15f7e07910b85f8a0437123e5fe174f71e761fc85e2aa2906954a80
+size 1247281
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001321.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001321.png
new file mode 100644
index 0000000000000000000000000000000000000000..b686917cf710adf5c16ebf86ffdc2f7af2d60fd9
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001321.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57b4f8487ac7eae179c32190e4017666c4e10d90b5075f4a03d1ea0bbc7b3a3c
+size 1465496
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001330.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001330.png
new file mode 100644
index 0000000000000000000000000000000000000000..ab2e46e9831527e60064f2216fddfed49b65572e
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001330.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47547bc563c624ac8ade81033126d80a5dfe8b98a3ba0b4524c96b2e36f7050b
+size 1531457
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001341.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001341.png
new file mode 100644
index 0000000000000000000000000000000000000000..b7a12a98dd2f7e72553d3588731b443296b8d236
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001341.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b24eed9cf6b6cf7a458d37ea128d100345707f3362f73f7f5cbf4589615787cb
+size 1581097
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001351.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001351.png
new file mode 100644
index 0000000000000000000000000000000000000000..b5ffdf522d64c0cea626b5151905a98b6ed26a7c
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001351.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6531c01bbc34a4d20d361da539d056bae1bb69154682cd50b76ccd7f09d24be5
+size 2922882
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001355.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001355.png
new file mode 100644
index 0000000000000000000000000000000000000000..112a21a64d1aa8ee027adf5f0ebd011f22e5097e
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001355.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bdcaf934b9098cddc46f18e5294bc4fc1a31331029300e4b6bb2abeadf5c3862
+size 2447802
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001364.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001364.png
new file mode 100644
index 0000000000000000000000000000000000000000..99ae086d27679d741ec73ae56c12ba7f0e62f549
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001364.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60750ff0879c6cbca019330b5e471c2ae19b1fdfb9c92453cef5b03d60120b84
+size 1535926
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001366.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001366.png
new file mode 100644
index 0000000000000000000000000000000000000000..fdb06e9ddc0d463afb965e794d47c1196462b249
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001366.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab9447b4fa6629c36a3e70bc4d4397211e082b2f25735262948386c93ee9a186
+size 1839587
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001368.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001368.png
new file mode 100644
index 0000000000000000000000000000000000000000..2a61b3413ec5c55bf3ae129d3532cf89f03905d5
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001368.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:130136cacc2ae8e31181804d3d46f433be4ea7331d8b6bf07ef2ead8253849f7
+size 1760505
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001390.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001390.png
new file mode 100644
index 0000000000000000000000000000000000000000..8965e6a25b5bd2608e6f939c35155810680dbcbf
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001390.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07a104d712378207aab56ff9735bd93e0d7eb9472d3d64aeb17a1e29517e68d5
+size 1142533
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001393.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001393.png
new file mode 100644
index 0000000000000000000000000000000000000000..31e81a8e118ae29e1799bcaa21057c0b1c8e3945
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001393.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:acfacbd7047f56b718924e11ccaa60ed0066c7daffa598d03309822bf25ff58f
+size 2403742
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001444.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001444.png
new file mode 100644
index 0000000000000000000000000000000000000000..1783e35753460cbca7cfb4d95559ebdcd312d546
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001444.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd146463c5f89f828b24d02bd129aafbe6960356c071881d572a023241d81f7e
+size 2126694
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001448.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001448.png
new file mode 100644
index 0000000000000000000000000000000000000000..f715471332a8e6730e689649b99e78d671750fdd
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001448.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ff3569be91a62a895c5fe5bdabdba1ba166924fca2e5507337437283d574171
+size 1411368
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001452.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001452.png
new file mode 100644
index 0000000000000000000000000000000000000000..04c64a335c935d55582c51684646cdd204b898c3
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001452.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e2d5a41732737dd58c3cb9f9259f59ccef7b1d5c0c96718f3efb407c42c6c6f
+size 1569597
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001457.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001457.png
new file mode 100644
index 0000000000000000000000000000000000000000..d996b573ec172405f635765e1d548efd2d2ce378
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001457.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c358e2197bcc1e67377e6b59da49742337e4c948100aa77f75ecd1076060d9c
+size 1962289
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001464.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001464.png
new file mode 100644
index 0000000000000000000000000000000000000000..5502fe8d9e015e36f49885f3f5c6c8d30f606dcf
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001464.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60c76569eb9ed201499bcc1ef1f56292f4fd9f404a7ee3881d1fd6aae0b53539
+size 1815851
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001484.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001484.png
new file mode 100644
index 0000000000000000000000000000000000000000..e74312a02b4e931fb16f5942eab6856065733607
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001484.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11102530b74690fd127ffba6df7e8b69b99726aacbf163f3e3d3752555fb3cc8
+size 1495764
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001488.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001488.png
new file mode 100644
index 0000000000000000000000000000000000000000..cbaebacf2b524f4a8c6b9f5da2728b324e06fb4d
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001488.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ccc224aa6f55cd1d4f77a3317a0e74e64a3ef721db28af17ef7e07063f61164e
+size 1817969
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001499.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001499.png
new file mode 100644
index 0000000000000000000000000000000000000000..3507d5878892bf3d410b98157b126d8e74949b32
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001499.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd5dec37a8c401cfaa46cb0950d70729375c0b574e2e359c256937ee1c88f52c
+size 1265530
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001505.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001505.png
new file mode 100644
index 0000000000000000000000000000000000000000..8fef44c2fb42715279190a40b338d8c67f15afb2
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001505.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:158ad03b7a45bb21f92cb7079dd9a0cb26df2dd008c998c83e34dcf3fd0ac548
+size 2111879
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001519.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001519.png
new file mode 100644
index 0000000000000000000000000000000000000000..39ccf15e709e00bd5410929014054a494bb1c784
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001519.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0fa35f68d84a2b53e39e6431811ca36da488bdb6688c6d3fc2484223250bf009
+size 1862990
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001522.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001522.png
new file mode 100644
index 0000000000000000000000000000000000000000..2ea4ef35da73cbcbbe6a39daf79b025ffe507783
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001522.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2bf556dbd5616a7c19866de38545fb7656bacf8d1cc1134979ab916900e0af07
+size 1492657
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001541.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001541.png
new file mode 100644
index 0000000000000000000000000000000000000000..e8fb3ec7cc3c3873e21d36a65cdb61babe6a17da
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001541.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:090f577b66d437744af322afdca034a20aa0a813123cad3c254eb1a65a81f923
+size 2283099
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001542.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001542.png
new file mode 100644
index 0000000000000000000000000000000000000000..ec7c931a6e0996efd60156a001a9dd415f930005
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001542.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ace13915a7b54f134ceba024dddbaeb853fc4c1f45a487a0140c90108ee6546
+size 1703520
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001549.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001549.png
new file mode 100644
index 0000000000000000000000000000000000000000..21725e233408c57f43b015282918fac123b65d15
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001549.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e535ed56d57a9e3c10b54aca09a01d0f6c895ca893605b6b8840e6cbef4371c6
+size 1500061
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001577.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001577.png
new file mode 100644
index 0000000000000000000000000000000000000000..761b5c0f12033136db8a34cb670982a077885dfe
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001577.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:182d83c7db3310c5abf425c18240c5e2ad24e7b83e3db1cec92d96899aa15292
+size 1515742
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001593.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001593.png
new file mode 100644
index 0000000000000000000000000000000000000000..8394f278fb2c0c36040e32d0fa9de8424fd6a5d6
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001593.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a3c88f760a79cc506f00e3828f256dce94516ddd994a232246cb3f0a56a78bd
+size 1632051
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001596.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001596.png
new file mode 100644
index 0000000000000000000000000000000000000000..d39d6b36eed8026676f35bfcebe7fe77dd72e1d8
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001596.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:911c6771448e06d967c6aacedae0bdbb26b56c1715623b514d69fccf7e0b5872
+size 1647149
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001612.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001612.png
new file mode 100644
index 0000000000000000000000000000000000000000..f1693e8aec2b6183385a46c897f3bd367a8bfffa
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001612.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b32704c3e11f845fb419e959a0b482cbfdc5e445cf2cee8fe84c61f819885dfb
+size 1559922
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001649.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001649.png
new file mode 100644
index 0000000000000000000000000000000000000000..8cc075a6557ccc35fc3636e388b7eef3adee6f00
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001649.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb66b364d59cf09b82b1743435b223f822ca52f1fe8227164179f993229ff0a2
+size 1549014
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001671.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001671.png
new file mode 100644
index 0000000000000000000000000000000000000000..6742eb96cce609798508f47093b88c73cea1b6c7
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001671.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6385fb6b31a25f6688af54b2408784b5769e03f29e95d808af31bda0c4bbcabb
+size 1595647
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001679.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001679.png
new file mode 100644
index 0000000000000000000000000000000000000000..86297a31cb49931fda2ab22757b697ef79374fe3
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001679.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:166d8bafbf18974cc7e8aa7132167e5e70b82105dbc56bf4be25f59448efb389
+size 1247218
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001689.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001689.png
new file mode 100644
index 0000000000000000000000000000000000000000..2863d0e89fc4bad211b7145ca78d59707f71d2f9
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001689.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d81e95392e48ffb6b602e51827da50de5c3fcd8887973418ac3acfb40f09908b
+size 2265721
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001701.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001701.png
new file mode 100644
index 0000000000000000000000000000000000000000..df08e644dd7949f2b2c4f6f8804621ebe2a1b8dc
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001701.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:085d28b435352cb10f9cb569b3f9d2cd416b57892da20348b8e278b245028512
+size 1594733
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001704.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001704.png
new file mode 100644
index 0000000000000000000000000000000000000000..3ef40702efeb0e9f39f016721b6cc30049496625
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001704.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0c19833de92426a0daf80301f11dc105983ce22e1f21be9426539c8c3a8fa1a
+size 1128975
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001709.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001709.png
new file mode 100644
index 0000000000000000000000000000000000000000..44a737778e932e303393f2422ad2b90eeee6579f
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001709.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1fb8f050de85c469ba9838caf6d27daf856d957555b0fd9b14a4870fb92b1ed8
+size 1691726
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001740.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001740.png
new file mode 100644
index 0000000000000000000000000000000000000000..63a4a38c16a35b0b021208bf93543d839c4358ed
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001740.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7bc6cd9eaeafba0d12f8e0892083b30588126dc42446159bd4abc87e4d4e858b
+size 2369989
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001744.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001744.png
new file mode 100644
index 0000000000000000000000000000000000000000..1bc2ea0c6d31433e9ee1e0d89184a05d4516c028
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001744.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69937501582123d9cb2e8d17fa365b9f4548e8ba62f70f0adb4e2b0dded5d304
+size 1811604
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001749.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001749.png
new file mode 100644
index 0000000000000000000000000000000000000000..db4953754e7d615c3cf0543edc9ea4018ebfadc7
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001749.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6801fad283398cd1d1c0e69d12b52f844d07aca120a98c8a343abc8dc1b6f201
+size 1396168
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001751.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001751.png
new file mode 100644
index 0000000000000000000000000000000000000000..5fb42f3b859e76e7ad3c1ef84c946fa5c319186e
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001751.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dae8caeb386716ad75e7f15f781b42e383df66c7653dbeb0ccfadc438c64c7c5
+size 1182259
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001760.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001760.png
new file mode 100644
index 0000000000000000000000000000000000000000..a18855bb61600e207d50f1af550216880e4b6f69
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001760.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71cff51a497bf61d3b5d8965d87d5cdd8c013210560b9889d66158cb2839b214
+size 1489583
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001762.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001762.png
new file mode 100644
index 0000000000000000000000000000000000000000..4a2dc021744535b091448a847b4c29eea09921eb
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001762.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:487a39bcd6009a5287b2d0a853fc0bf065dd84f9fc4fbe348ee214d8892897a5
+size 1603477
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001781.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001781.png
new file mode 100644
index 0000000000000000000000000000000000000000..2af8e926c0d7d3c1b768d92adb7b3c017a1099b3
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001781.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:433fff69a25bda64914ff3ba803f5ec4276f1b00ce1c6cd261ff2ed60367d23d
+size 1857930
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001791.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001791.png
new file mode 100644
index 0000000000000000000000000000000000000000..00e0d99c3da7939cb49bd4e78f3d382b2c75d0f2
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001791.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d039508cd47ecdfa75b07be17e917b6d881b05490504ad2762ef6bcd49f4bb9f
+size 1077459
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001843.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001843.png
new file mode 100644
index 0000000000000000000000000000000000000000..5609013ce63f83bbcfbcbb926961b6585d1afcce
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001843.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ea4341f4bcb7dc00241abb4e537e0bdd40f9007c9fc61d07e747752f32f6469
+size 1303932
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001854.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001854.png
new file mode 100644
index 0000000000000000000000000000000000000000..0d27ab41b67bde351e0a1247ea0e1fb9ca04aa3b
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001854.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc468340853d421cb8b79df9d5416a6772f9975233528f84258fed3e0f475c76
+size 1338534
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001856.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001856.png
new file mode 100644
index 0000000000000000000000000000000000000000..54c14a28a8b71928d4ba901f3878e0b9b4f5d395
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001856.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac6cb538b372a24ec97bdd813eb991ec1da82789f8f300a84d97715d051b23f1
+size 1415595
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001867.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001867.png
new file mode 100644
index 0000000000000000000000000000000000000000..d7ab58078a515dd2542aa770c7f2245474684323
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001867.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1026ee125d9ae4e22f08e579d0f671d327b922d0bbefab28032c1241663cb347
+size 1568125
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001884.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001884.png
new file mode 100644
index 0000000000000000000000000000000000000000..e6944219c8583086e2a20f83c47e93d9fc001a9e
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001884.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b76e07a3011cc015f3b18b5c1d3616e32057953bba58e0063284cff657fc017
+size 1461638
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001896.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001896.png
new file mode 100644
index 0000000000000000000000000000000000000000..172225b8bfaf0687feb086604e1bf22e2c3e093b
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001896.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21fa5b0df5f24d595129550bedfa627577f5b1e1f86f133c00750e8502f45939
+size 1224894
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001907.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001907.png
new file mode 100644
index 0000000000000000000000000000000000000000..1e694f2b46b75d4afad1f40b6774760593d69d7d
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001907.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d896ad7fbc854e0b5134885a48b04fa4bf8cba51a0356189a4d16525c8da133
+size 1206724
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001910.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001910.png
new file mode 100644
index 0000000000000000000000000000000000000000..311c9e67346ae6177499e1a16bc02feee2578759
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001910.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c470ff63e8e6de40f399abf4b32228db37e368397ec84503bb488ed4108239ed
+size 1761674
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001919.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001919.png
new file mode 100644
index 0000000000000000000000000000000000000000..d5ef6729241e8d7d47ad1a154be960185009f849
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001919.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5bd761b4b1824a60a5f604eb9ae8f049228da69321cf5b5956868cfa8b03e87
+size 1513172
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001940.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001940.png
new file mode 100644
index 0000000000000000000000000000000000000000..fed541582e36f877a2bd822114ebab8d3be9e6a2
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_001940.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d58aea4da382dee3b926c1f6e8365f6206f4468eaf60caf74e76f9aa9b703479
+size 1185835
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002008.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002008.png
new file mode 100644
index 0000000000000000000000000000000000000000..312e13ee97079bc6f90f3bbb85de0ceb86a8e6d3
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002008.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3714297f352fe6d42b13c49884453d933605a0f2db419183b687c7277eb861e1
+size 1406442
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002062.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002062.png
new file mode 100644
index 0000000000000000000000000000000000000000..49c157db67179bb95c934dac65714882d7f603c5
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002062.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1eabc24dc5ebcab47ebc4461da9bf0d819ed1a3ad5271cd6406ce7a82bcb0c77
+size 1660511
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002077.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002077.png
new file mode 100644
index 0000000000000000000000000000000000000000..b030fb63cfe59cef788212e5df8c2abe2c2951a9
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002077.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20ad3f55bb1114035701c1f5b967880aba7159b0e631a17380443b4a2767a8d9
+size 1812464
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002089.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002089.png
new file mode 100644
index 0000000000000000000000000000000000000000..1fea4ae7456493fd730dd35a7e385e98ca3a9afb
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002089.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d90609e8b335972d589f3e3d36042e4e4092573694efe52e00d79320f418da83
+size 1231908
diff --git a/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002090.png b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002090.png
new file mode 100644
index 0000000000000000000000000000000000000000..5078dd25a64a5dc7c45599ef897b5c48da9edfc7
--- /dev/null
+++ b/image_12000triple_dpo_rwr_1e-4_250/fintune_2epoch_002090.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f184ea3af082f0f05a9a5dad05314f9468fe9e4076b39a648f89015753e60d35
+size 1424323
diff --git a/trained-sd3_5[dpo_and_rwr]_[chengyiw-w,l-l]_[p=05]_[batch_size4]_[beta250]_[1e-4]/train_dpo_lora_sd3_rwr.py b/trained-sd3_5[dpo_and_rwr]_[chengyiw-w,l-l]_[p=05]_[batch_size4]_[beta250]_[1e-4]/train_dpo_lora_sd3_rwr.py
new file mode 100644
index 0000000000000000000000000000000000000000..e5c99fe72975a370c20fbe50c676e69095fa335e
--- /dev/null
+++ b/trained-sd3_5[dpo_and_rwr]_[chengyiw-w,l-l]_[p=05]_[batch_size4]_[beta250]_[1e-4]/train_dpo_lora_sd3_rwr.py
@@ -0,0 +1,2268 @@
+#!/usr/bin/env python
+# coding=utf-8
+# Copyright 2025 The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+
+import argparse
+import copy
+import itertools
+import logging
+import math
+import os
+import random
+import shutil
+import warnings
+from contextlib import nullcontext
+from pathlib import Path
+import pandas as pd
+from collections import defaultdict
+
+import numpy as np
+import torch
+import torch.utils.checkpoint
+import transformers
+from accelerate import Accelerator, DistributedType
+from accelerate.logging import get_logger
+from accelerate.utils import DistributedDataParallelKwargs, ProjectConfiguration, set_seed
+from huggingface_hub import create_repo, upload_folder
+from huggingface_hub.utils import insecure_hashlib
+from peft import LoraConfig, set_peft_model_state_dict
+from peft.utils import get_peft_model_state_dict
+from PIL import Image
+from PIL.ImageOps import exif_transpose
+from torch.utils.data import Dataset
+from torchvision import transforms
+from torchvision.transforms.functional import crop
+from tqdm.auto import tqdm
+from transformers import CLIPTokenizer, PretrainedConfig, T5TokenizerFast
+import torch.nn.functional as F
+
+import diffusers
+from diffusers import (
+ AutoencoderKL,
+ FlowMatchEulerDiscreteScheduler,
+ SD3Transformer2DModel,
+ StableDiffusion3Pipeline,
+)
+from diffusers.optimization import get_scheduler
+from diffusers.training_utils import (
+ _set_state_dict_into_text_encoder,
+ cast_training_params,
+ compute_density_for_timestep_sampling,
+ compute_loss_weighting_for_sd3,
+ free_memory,
+)
+from diffusers.utils import (
+ check_min_version,
+ convert_unet_state_dict_to_peft,
+ is_wandb_available,
+)
+from diffusers.utils.hub_utils import load_or_create_model_card, populate_model_card
+from diffusers.utils.torch_utils import is_compiled_module
+
+
+if is_wandb_available():
+ import wandb
+
+os.environ["WANDB_API_KEY"] = 'c3c7dc2e7a43cc9e0b4cc8e913d363077af04ab2'
+os.environ["WANDB_MODE"] = "offline"
+# Will error if the minimal version of diffusers is not installed. Remove at your own risks.
+check_min_version("0.33.0.dev0")
+
+logger = get_logger(__name__)
+
+DATASET_NAME_MAPPING = {
+ # "refl": ("image", "text"),
+}
+
+#将模型的相关信息保存到readme.md文件中
+def save_model_card(
+ repo_id: str,
+ images=None,
+ base_model: str = None,
+ train_text_encoder=False,
+ instance_prompt=None,
+ validation_prompts=None,
+ repo_folder=None,
+):
+ if "large" in base_model:
+ model_variant = "SD3.5-Large"
+ license_url = "https://huggingface.co/stabilityai/stable-diffusion-3.5-large/blob/main/LICENSE.md"
+ variant_tags = ["sd3.5-large", "sd3.5", "sd3.5-diffusers"]
+ else:
+ model_variant = "SD3"
+ license_url = "https://huggingface.co/stabilityai/stable-diffusion-3-medium/blob/main/LICENSE.md"
+ variant_tags = ["sd3", "sd3-diffusers"]
+
+ widget_dict = []
+ if images is not None:
+ for i, image in enumerate(images):
+ image.save(os.path.join(repo_folder, f"image_{i}.png"))
+ widget_dict.append(
+ {"text": validation_prompts if validation_prompts else " ", "output": {"url": f"image_{i}.png"}}
+ )#将验证的prompt和其对应的图片的信息一起保存下来
+
+ model_description = f"""
+# {model_variant} DreamBooth LoRA - {repo_id}
+
+
+
+## Model description
+
+These are {repo_id} DreamBooth LoRA weights for {base_model}.
+
+The weights were trained using [DreamBooth](https://dreambooth.github.io/) with the [SD3 diffusers trainer](https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/README_sd3.md).
+
+Was LoRA for the text encoder enabled? {train_text_encoder}.
+
+## Trigger words
+
+You should use `{instance_prompt}` to trigger the image generation.
+
+## Download model
+
+[Download the *.safetensors LoRA]({repo_id}/tree/main) in the Files & versions tab.
+
+## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
+
+```py
+from diffusers import AutoPipelineForText2Image
+import torch
+pipeline = AutoPipelineForText2Image.from_pretrained({base_model}, torch_dtype=torch.float16).to('cuda')
+pipeline.load_lora_weights('{repo_id}', weight_name='pytorch_lora_weights.safetensors')
+image = pipeline('{validation_prompts }').images[0]
+```
+
+### Use it with UIs such as AUTOMATIC1111, Comfy UI, SD.Next, Invoke
+
+- **LoRA**: download **[`diffusers_lora_weights.safetensors` here 💾](/{repo_id}/blob/main/diffusers_lora_weights.safetensors)**.
+ - Rename it and place it on your `models/Lora` folder.
+ - On AUTOMATIC1111, load the LoRA by adding `` to your prompt. On ComfyUI just [load it as a regular LoRA](https://comfyanonymous.github.io/ComfyUI_examples/lora/).
+
+For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
+
+## License
+
+Please adhere to the licensing terms as described [here]({license_url}).
+"""
+ model_card = load_or_create_model_card(
+ repo_id_or_path=repo_id,
+ from_training=True,
+ license="other",
+ base_model=base_model,
+ model_description=model_description,
+ widget=widget_dict,
+ )
+ tags = [
+ "text-to-image",
+ "diffusers-training",
+ "diffusers",
+ "lora",
+ "template:sd-lora",
+ ]
+
+ tags += variant_tags
+
+ model_card = populate_model_card(model_card, tags=tags)
+ model_card.save(os.path.join(repo_folder, "README.md"))
+
+#加载文本编码器
+def load_text_encoders(class_one, class_two, class_three):
+ text_encoder_one = class_one.from_pretrained(
+ args.pretrained_model_name_or_path, subfolder="text_encoder", revision=args.revision, variant=args.variant
+ )
+ text_encoder_two = class_two.from_pretrained(
+ args.pretrained_model_name_or_path, subfolder="text_encoder_2", revision=args.revision, variant=args.variant
+ )
+ text_encoder_three = class_three.from_pretrained(
+ args.pretrained_model_name_or_path, subfolder="text_encoder_3", revision=args.revision, variant=args.variant
+ )
+ return text_encoder_one, text_encoder_two, text_encoder_three
+
+#将每次检测的照片保存成日志的形式，并上传到wandb上面
+def log_validation(
+ pipeline,
+ args,
+ accelerator,
+ pipeline_args, # 现在是一个 prompt 配置组成的 list
+ global_step,
+ torch_dtype,
+ is_final_validation=False,
+):
+ logger.info(
+ f"Running validation... \n Generating {len(pipeline_args)} images with prompts:"
+ f" {[args['prompt'] for args in pipeline_args]}."
+ )
+ pipeline = pipeline.to(accelerator.device)
+ pipeline.set_progress_bar_config(disable=True)
+
+ autocast_ctx = nullcontext()
+ generator = torch.Generator(device=accelerator.device).manual_seed(args.seed) if args.seed is not None else None
+
+ # generate one image for each prompt dict in pipeline_args
+ images = []
+ with autocast_ctx:
+ for single_args in pipeline_args:
+ image = pipeline(**single_args, generator=generator).images[0]
+ images.append(image)
+
+ # Logging to trackers
+ for tracker in accelerator.trackers:
+ phase_name = "test" if is_final_validation else "validation"
+ if tracker.name == "tensorboard":
+ np_images = np.stack([np.asarray(img) for img in images])
+ tracker.writer.add_images(phase_name, np_images, global_step, dataformats="NHWC")
+ elif tracker.name == "wandb":
+ tracker.log({
+ phase_name: [
+ wandb.Image(image, caption=f"{i}: {args['prompt']}")
+ for i, (image, args) in enumerate(zip(images, pipeline_args))
+ ]
+ })
+
+ del pipeline
+ free_memory()
+ return images
+
+
+#从所给的模型名称或者是路径中，加载模型配置，从而导入相应模型的类
+def import_model_class_from_model_name_or_path(
+ pretrained_model_name_or_path: str, revision: str, subfolder: str = "text_encoder"
+):
+ text_encoder_config = PretrainedConfig.from_pretrained(
+ pretrained_model_name_or_path, subfolder=subfolder, revision=revision
+ )
+ model_class = text_encoder_config.architectures[0]
+ if model_class == "CLIPTextModelWithProjection":
+ from transformers import CLIPTextModelWithProjection
+
+ return CLIPTextModelWithProjection
+ elif model_class == "T5EncoderModel":
+ from transformers import T5EncoderModel
+
+ return T5EncoderModel
+ else:
+ raise ValueError(f"{model_class} is not supported.")
+
+
+#设置和解析命令行参数
+def parse_args(input_args=None):
+ parser = argparse.ArgumentParser(description="Simple example of a training script.")
+ parser.add_argument(
+ "--pretrained_model_name_or_path",
+ type=str,
+ default=None,
+ required=True,
+ help="Path to pretrained model or model identifier from huggingface.co/models.",
+ )
+ parser.add_argument(
+ "--revision",
+ type=str,
+ default=None,
+ required=False,
+ help="Revision of pretrained model identifier from huggingface.co/models.",
+ )#如果不提供这个参数的话：
+# 如果你从的是一个普通的 checkpoint 目录（例如你本地保存的微调模型），没有版本分支概念，那根本不会有任何影响。
+# 如果你从的是 Hugging Face 上的模型仓库，比如 "stabilityai/stable-diffusion-3"，那么默认加载的是该 repo 的默认分支。
+ parser.add_argument(
+ "--variant",
+ type=str,
+ default=None,
+ help="Variant of the model files of the pretrained model identifier from huggingface.co/models, 'e.g.' fp16",
+ )#--variant 指的是预训练模型的不同文件变种（variant）
+ #预训练模型通常托管在 Hugging Face Hub（huggingface.co/models）上，有时候一个模型会有不同的版本，
+ #例如 fp16 和 fp32 版本，或者有不同的精度（如 bfloat16），这时候就需要使用 --variant 参数来指定。
+ #如果不指定的话，那就用默认的（通常是 full-precision 版本）
+ parser.add_argument(
+ "--dataset_name",
+ type=str,
+ default=None,
+ help=(
+ "The name of the Dataset (from the HuggingFace hub) containing the training data of instance images (could be your own, possibly private,"
+ " dataset). It can also be a path pointing to a local copy of a dataset in your filesystem,"
+ " or to a folder containing files that 🤗 Datasets can understand."
+ ),
+ )
+ parser.add_argument(
+ "--dataset_config_name",
+ type=str,
+ default=None,
+ help="The config of the Dataset, leave as None if there's only one config.",
+ )
+
+ parser.add_argument(
+ "--cache_dir",
+ type=str,
+ default=None,
+ help="The directory where the downloaded models and datasets will be stored.",
+ )
+ parser.add_argument("--repeats", type=int, default=1, help="How many times to repeat the training data.")
+
+ # parser.add_argument(
+ # "--reward_file",
+ # type=str,
+ # default=None,
+ # help="The file stored the reward model scores of all images in database.",
+ # )
+
+ parser.add_argument(
+ "--image_column",
+ type=str,
+ default="image",
+ help="The column of the dataset containing the target image. By "
+ "default, the standard Image Dataset maps out 'file_name' "
+ "to 'image'.",
+ )
+ parser.add_argument(
+ "--caption_column",
+ type=str,
+ default=None,
+ help="The column of the dataset containing the instance prompt for each image",
+ )
+
+
+ parser.add_argument(
+ "--max_sequence_length",
+ type=int,
+ default=512,
+ help="Maximum sequence length to use with with the T5 text encoder",
+ )
+ parser.add_argument(
+ "--validation_prompts",
+ type=str,
+ nargs="+",#这个是表示接受一个值或多个值，并不是说参数之间用加号连接
+ default=None,
+ help="A prompt that is used during validation to verify that the model is learning. The validation is happening at each `--checkpointing_steps`."
+ )
+ parser.add_argument(
+ "--num_validation_images",
+ type=int,
+ default=4,
+ help="Number of images that should be generated during validation with `validation_prompts`.",
+ )#配合 --sample_batch_size 使用：一般会循环调用若干次 sampling 来生成足够图像
+ parser.add_argument(
+ "--validation_steps",
+ type=int,
+ default=1,
+ help=(
+ "Run dreambooth validation every X epochs. Dreambooth validation consists of running the prompt"
+ " `args.validation_prompts` multiple times: `args.num_validation_images`."
+ ),
+ )
+ parser.add_argument(
+ "--rank",
+ type=int,
+ default=768,
+ help=("The dimension of the LoRA update matrices."),
+ )
+ parser.add_argument(
+ "--beta_dpo",
+ type=int,
+ default=2500,
+ help="DPO KL Divergence penalty.",
+ )
+ #LoRA 会用两个 rank 更低的矩阵（通常是 A 和 B）来近似训练参数的变化，而不直接修改原始模型参数。
+ #rank 决定这两个矩阵的维度，rank 越大，表示 LoRA 模拟的能力越强（也越耗显存）。
+ #通常选择： 4、8、16 比较常见。
+
+
+
+ parser.add_argument(
+ "--output_dir",
+ type=str,
+ default="sd3-dreambooth",
+ help="The output directory where the model predictions and checkpoints will be written.",
+ )
+ parser.add_argument("--seed", type=int, default=None, help="A seed for reproducible training.")
+ parser.add_argument(
+ "--resolution",
+ type=int,
+ default=1024,
+ help=(
+ "The resolution for input images, all the images in the train/validation dataset will be resized to this"
+ " resolution"
+ ),
+ )
+ parser.add_argument(
+ "--center_crop",
+ default=True,
+ action="store_true",
+ help=(
+ "Whether to center crop the input images to the resolution. If not set, the images will be randomly"
+ " cropped. The images will be resized to the resolution first before cropping."
+ ),
+ )
+ parser.add_argument(
+ "--random_flip",
+ action="store_true",
+ help="whether to randomly flip images horizontally",
+ )
+ parser.add_argument(
+ "--train_text_encoder",
+ action="store_true",
+ help="Whether to train the text encoder (clip text encoders only). If set, the text encoder should be float32 precision.",
+ )
+
+ parser.add_argument(
+ "--train_batch_size", type=int, default=4, help="Batch size (per device) for the training dataloader."
+ )
+ parser.add_argument(
+ "--sample_batch_size", type=int, default=4, help="Batch size (per device) for sampling images."
+ )#这个参数控制在训练或验证过程中，生成图像（inference）时，每张 GPU 同时采样几张图片。
+ #采样过程不需要反向传播（即不占用梯度空间），适用于如 validation 生成、推理、展示效果等场景。
+ parser.add_argument("--num_train_epochs", type=int, default=100)
+ parser.add_argument(
+ "--max_train_steps",
+ type=int,
+ default=None,
+ help="Total number of training steps to perform. If provided, overrides num_train_epochs.",
+ )
+ parser.add_argument(
+ "--checkpointing_steps",
+ type=int,
+ default=500,
+ help=(
+ "Save a checkpoint of the training state every X updates. These checkpoints can be used both as final"
+ " checkpoints in case they are better than the last checkpoint, and are also suitable for resuming"
+ " training using `--resume_from_checkpoint`."
+ ),
+ )
+ parser.add_argument(
+ "--checkpoints_total_limit",
+ type=int,
+ default=None,
+ help=("Max number of checkpoints to store."),
+ )
+ parser.add_argument(
+ "--resume_from_checkpoint",
+ type=str,
+ default=None,
+ help=(
+ "Whether training should be resumed from a previous checkpoint. Use a path saved by"
+ ' `--checkpointing_steps`, or `"latest"` to automatically select the last available checkpoint.'
+ ),
+ )
+ parser.add_argument(
+ "--gradient_accumulation_steps",
+ type=int,
+ default=1,
+ help="Number of updates steps to accumulate before performing a backward/update pass.",
+ )
+ parser.add_argument(
+ "--gradient_checkpointing",
+ action="store_true",
+ help="Whether or not to use gradient checkpointing to save memory at the expense of slower backward pass.",
+ )
+ parser.add_argument(
+ "--learning_rate",
+ type=float,
+ default=1e-4,
+ help="Initial learning rate (after the potential warmup period) to use.",
+ )
+
+ parser.add_argument(
+ "--text_encoder_lr",
+ type=float,
+ default=5e-6,
+ help="Text encoder learning rate to use.",
+ )
+ parser.add_argument(
+ "--scale_lr",
+ action="store_true",
+ default=False,
+ help="Scale the learning rate by the number of GPUs, gradient accumulation steps, and batch size.",
+ )
+ parser.add_argument(
+ "--lr_scheduler",
+ type=str,
+ default="constant",
+ help=(
+ 'The scheduler type to use. Choose between ["linear", "cosine", "cosine_with_restarts", "polynomial",'
+ ' "constant", "constant_with_warmup"]'
+ ),
+ )
+ parser.add_argument(
+ "--lr_warmup_steps", type=int, default=500, help="Number of steps for the warmup in the lr scheduler."
+ )
+ parser.add_argument(
+ "--lr_num_cycles",
+ type=int,
+ default=1,
+ help="Number of hard resets of the lr in cosine_with_restarts scheduler.",
+ )
+ # 这个调度器（cosine_with_restarts）的工作机制是：
+ # 学习率（learning rate）按余弦函数的曲线逐渐降低。
+ # 每过一段时间，会突然“重置”学习率，从高值重新开始一个新的下降周期，以跳出局部最优。
+ #lr_num_cycles就是用来控制在整个训练过程中，要进行几次这样的学习率下降 → 重置 的周期。
+ # 为什么使用 restart？
+ # 模拟“多次训练”的感觉
+ # 避免模型陷入局部最优
+ # 提高收敛稳定性和泛化能力
+ parser.add_argument("--lr_power", type=float, default=1.0, help="Power factor of the polynomial scheduler.")
+ # 这个参数控制学习率下降的速率。
+ # 如果设置为 1.0，学习率会按线性速率下降。
+ # 如果设置为 2.0，学习率会按平方速率下降。
+ # 如果设置为 3.0，学习率会按立方速率下降。
+ # 通常使用 1.0 或 2.0。
+
+ parser.add_argument(
+ "--dataloader_num_workers",
+ type=int,
+ default=0,
+ help=(
+ "Number of subprocesses to use for data loading. 0 means that the data will be loaded in the main process."
+ ),
+ )
+ parser.add_argument(
+ "--weighting_scheme",
+ type=str,
+ default="logit_normal",
+ choices=["sigma_sqrt", "logit_normal", "mode", "cosmap"],
+ )#--weighting_scheme 就是 决定你训练时该偏爱哪些样本 的方式，每种策略都有它“偏爱的对象”
+ #sigma_sqrt:
+ #逻辑： 权重 = sqrt(σ)，σ 是某种噪声或不确定性
+ # 权重与噪声方差 $\sigma$ 的平方根有关，强调不同噪声强度下的训练平衡，噪声越大权重越大。
+ # 适用于： 噪声方差变化较大的数据集，如多模态数据。
+ # 优点： 平衡不同噪声水平下的训练，避免某些样本权重过大。
+ # 缺点： 可能需要手动调整参数，以适应不同数据集的噪声特性。
+ #logit_normal:
+ #逻辑： 权重 = logit(reward)，σ 是某种噪声或不确定性
+ #把 reward 想成“考试得分”，先经过一个 S 型函数（像 sigmoid），再转成权重。
+ #logit_mean 和 logit_std 控制了这个 S 曲线的中点和陡峭程度
+ #mode:
+ #逻辑： reward 的众数（最常见的高分）作为中心，越接近就越重要
+ # 适用于： 奖励分布较集中（如二元分类）的数据集。
+ # 优点： 强调常见高分样本，有助于模型学习。
+ # 缺点： 可能需要手动调整参数，以适应不同数据集的奖励分布。
+ #cosmap:
+ #逻辑： 用余弦函数把 reward 映射成 [0, 1] 的权重
+ #形象比喻： 一个像海浪一样起伏的权重函数，中间 reward 得到最高权重，两边低。
+ # 更平滑的加权方式，适合 reward 不是特别尖锐地分布的情况。
+
+ parser.add_argument(
+ "--logit_mean", type=float, default=0.0, help="mean to use when using the `'logit_normal'` weighting scheme."
+ )#--logit_mean 和 --logit_std 是 logit_normal 策略的参数，用于控制样本的权重分布。
+ # 它们决定了样本权重在 logit 空间中的均值和标准差。
+ # 均值（mean）：样本权重在 logit 空间中的中心位置。
+ # 标准差（std）：样本权重在 logit 空间中的离散程度。
+ # 这两个参数共同决定了样本权重在 logit 空间中的分布形状。
+ # 通常，均值设置为 0.0，标准差设置为 1.0。
+
+ parser.add_argument(
+ "--logit_std", type=float, default=1.0, help="std to use when using the `'logit_normal'` weighting scheme."
+ )
+ parser.add_argument(
+ "--mode_scale",
+ type=float,
+ default=1.29,
+ help="Scale of mode weighting scheme. Only effective when using the `'mode'` as the `weighting_scheme`.",
+ )#--mode_scale 是 mode 策略的参数，用于控制样本的权重分布。
+ #用来放大或缩小以众数为核心的加权范围，数值越大，对高 reward 样本越偏向。
+
+ parser.add_argument(
+ "--precondition_outputs",
+ type=int,
+ default=1,
+ help="Flag indicating if we are preconditioning the model outputs or not as done in EDM. This affects how "
+ "model `target` is calculated.",
+ )#这个参数控制是否对模型输出进行预处理（preconditioning），这通常出现在 EDM（Elucidated Diffusion Models） 中
+ #如果为 0：使用原始的 target 计算损失。
+ #如果为 1：使用按照EDM的方法调整后的 target 计算损失。
+ parser.add_argument(
+ "--optimizer",
+ type=str,
+ default="AdamW",
+ help=('The optimizer type to use. Choose between ["AdamW", "prodigy"]'),
+ )
+
+ parser.add_argument(
+ "--use_8bit_adam",
+ action="store_true",
+ help="Whether or not to use 8-bit Adam from bitsandbytes. Ignored if optimizer is not set to AdamW",
+ )#Adam（Adaptive Moment Estimation）是一种自适应学习率的优化算法，是 SGD（随机梯度下降）的改进版本
+ #结合了 Momentum（动量） 和 RMSProp（自适应学习率）,适合大部分深度学习任务，收敛很快
+ #8-bit Adam（由 bitsandbytes 库实现）将部分计算压缩到 8-bit 整数，能够减少内存占用，提升计算速度
+
+ # Prodigy 是一种新型的自适应优化器，可以看作是对 Adam 优化器的改进，主要目标是：
+ # 收敛更快：相比 Adam，在一些任务中它能更快收敛。
+ # 更稳定的训练过程：对学习率和权重衰减的响应更温和。
+ # 自带一些 bias correction、预热保护机制（warm-up safeguard），不容易在训练初期出现震荡。
+ # 它的核心思想是：
+ # 不仅跟踪一阶（梯度）和二阶（梯度平方）动量，还使用了额外的β3 系数来平滑地控制步长。
+ # 与 Adam 不同的是，Prodigy 还尝试根据历史梯度自动调整步长，不仅仅依赖学习率。
+ # ✅ 简单来说：Prodigy 是为了在保持 Adam 优点的同时提高效率和稳定性而设计的优化器。
+
+ parser.add_argument("--adam_beta1", type=float, default=0.9, help="The beta1 parameter for the Adam and Prodigy optimizers.")
+ # Adam 和 Prodigy 优化器中的 β1 参数，控制一阶矩估计的指数衰减率，默认是 0.9。
+
+ parser.add_argument("--adam_beta2", type=float, default=0.999, help="The beta2 parameter for the Adam and Prodigy optimizers.")
+ # Adam 和 Prodigy 优化器中的 β2 参数，控制二阶矩估计的指数衰减率，默认是 0.999。
+
+ parser.add_argument(
+ "--prodigy_beta3",
+ type=float,
+ default=None,
+ help="coefficients for computing the Prodigy stepsize using running averages. If set to None, "
+ "uses the value of square root of beta2. Ignored if optimizer is adamW",
+ )
+ # Prodigy 优化器中用于计算步长的 β3 系数，默认为 sqrt(beta2)。如果不使用 Prodigy 优化器将被忽略。
+
+ parser.add_argument("--prodigy_decouple", type=bool, default=True, help="Use AdamW style decoupled weight decay")
+ # 是否采用 AdamW 式的权重衰减方式（将权重衰减与梯度更新分离），通常能提升训练稳定性。
+
+ parser.add_argument("--adam_weight_decay", type=float, default=1e-04, help="Weight decay to use for unet params")
+ # 应用于 UNet 参数的权重衰减系数，有助于防止过拟合，默认值为 1e-4。
+
+ parser.add_argument("--adam_weight_decay_text_encoder", type=float, default=1e-03, help="Weight decay to use for text_encoder")
+ # 应用于 text_encoder 参数的权重衰减系数，默认值为 1e-3。
+
+ parser.add_argument(
+ "--lora_layers",
+ type=str,
+ default=None,
+ help=(
+ "The transformer block layers to apply LoRA training on. Please specify the layers in a comma seperated string."
+ "For examples refer to https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/README_SD3.md"
+ ),
+ )
+ # 指定在哪些 transformer 层上应用 LoRA 微调，多个层用逗号分隔，例如："q_proj,k_proj,v_proj"。
+
+ parser.add_argument(
+ "--lora_blocks",
+ type=str,
+ default=None,
+ help=(
+ "The transformer blocks to apply LoRA training on. Please specify the block numbers in a comma seperated manner."
+ 'E.g. - "--lora_blocks 12,30" will result in lora training of transformer blocks 12 and 30. For more examples refer to https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/README_SD3.md'
+ ),
+ )
+ # 指定要进行 LoRA 微调的 transformer block 编号，多个编号用逗号分隔，例如 "12,30"。
+ # --lora_blocks 控制在哪几层 transformer 中使用 LoRA
+ # --lora_layers 控制在这些层中 LoRA 作用于哪些子模块
+ #相当于是--lora_layers是比--lora_blocks更细粒度的控制
+ # 其实当时运行的时候虽然--lora_blocks和--lora_layers都设置为None，但是在这种情况下还是会使用lora参数的，只不过是默认是全部层都使用LoRA，这部分是perf库自动控制的
+
+ parser.add_argument(
+ "--adam_epsilon",
+ type=float,
+ default=1e-08,
+ help="Epsilon value for the Adam optimizer and Prodigy optimizers.",
+ )
+ # Adam 和 Prodigy 优化器中的 ε 值，用于防止除以零或数值不稳定，默认值为 1e-8。
+
+ parser.add_argument(
+ "--prodigy_use_bias_correction",
+ type=bool,
+ default=True,
+ help="Turn on Adam's bias correction. True by default. Ignored if optimizer is adamW",
+ )
+ # 是否开启 Adam 风格的偏差修正，能提升早期训练效果。仅在使用 Prodigy 优化器时生效。
+
+ parser.add_argument(
+ "--prodigy_safeguard_warmup",
+ type=bool,
+ default=True,
+ help="Remove lr from the denominator of D estimate to avoid issues during warm-up stage. True by default. "
+ "Ignored if optimizer is adamW",
+ )
+ # 是否在 warm-up 阶段避免除以学习率来稳定训练过程，仅在 Prodigy 优化器中有效。
+
+ parser.add_argument("--max_grad_norm", default=1.0, type=float, help="Max gradient norm.")
+ parser.add_argument("--push_to_hub", action="store_true", help="Whether or not to push the model to the Hub.")
+ parser.add_argument("--hub_token", type=str, default=None, help="The token to use to push to the Model Hub.")
+ parser.add_argument(
+ "--hub_model_id",
+ type=str,
+ default=None,
+ help="The name of the repository to keep in sync with the local `output_dir`.",
+ )
+ parser.add_argument(
+ "--logging_dir",
+ type=str,
+ default="logs",
+ help=(
+ "[TensorBoard](https://www.tensorflow.org/tensorboard) log directory. Will default to"
+ " *output_dir/runs/**CURRENT_DATETIME_HOSTNAME***."
+ ),
+ )
+ parser.add_argument(
+ "--allow_tf32",
+ action="store_true",
+ help=(
+ "Whether or not to allow TF32 on Ampere GPUs. Can be used to speed up training. For more information, see"
+ " https://pytorch.org/docs/stable/notes/cuda.html#tensorfloat-32-tf32-on-ampere-devices"
+ ),
+ )#--allow_tf32：是否允许 TF32 加速（仅适用于 Ampere 架构 GPU，如 A100、3090）TF32 计算比 FP32 快，但可能降低精度。
+ parser.add_argument(
+ "--cache_latents",
+ action="store_true",
+ default=False,
+ help="Cache the VAE latents",
+ )#--cache_latents：是否缓存 VAE 的潜在变量（latents）。
+ # 缓存可以加速训练，但会增加内存使用。
+ # 如果训练过程中出现内存不足，可以考虑关闭缓存。
+ # 如果训练稳定，可以保持默认值。
+ parser.add_argument(
+ "--report_to",
+ type=str,
+ default="tensorboard",
+ help=(
+ 'The integration to report the results and logs to. Supported platforms are `"tensorboard"`'
+ ' (default), `"wandb"` and `"comet_ml"`. Use `"all"` to report to all integrations.'
+ ),
+ )
+ parser.add_argument(
+ "--mixed_precision",
+ type=str,
+ default=None,
+ choices=["no", "fp16", "bf16"],
+ help=(
+ "Whether to use mixed precision. Choose between fp16 and bf16 (bfloat16). Bf16 requires PyTorch >="
+ " 1.10.and an Nvidia Ampere GPU. Default to the value of accelerate config of the current system or the"
+ " flag passed with the `accelerate.launch` command. Use this argument to override the accelerate config."
+ ),
+ )
+ parser.add_argument(
+ "--upcast_before_saving",
+ action="store_true",
+ default=False,
+ help=(
+ "Whether to upcast the trained transformer layers to float32 before saving (at the end of training). "
+ "Defaults to precision dtype used for training to save memory"
+ ),
+ )
+ # 是否在保存模型前将 transformer 层转换为 float32 精度（默认不转换，即保持训练时的精度）。
+ # 这样做可以避免低精度导致的数值不稳定，但会增加模型大小。
+
+ # Personal usage arguments
+ parser.add_argument(
+ "--apply_pre_loss", action="store_true", help="Whether or not to apply pretrained loss."
+ )
+ parser.add_argument(
+ "--apply_reward_loss", action="store_true", help="Whether or not to apply reward loss."
+ )
+ # parser.add_argument(
+ # "--image_reward_version",
+ # type=str,
+ # required=True,
+ # help=(
+ # "The version of ImageReward to load with. This could be a downloadable version, i.e. ImageReward-v1.0"
+ # "or a file path."),
+ # )
+ parser.add_argument(
+ "--save_only_one_ckpt",
+ action="store_true",
+ help="If given, then it only stores one checkpoint through the whole training, the one with the best training loss."
+ "This is useful to manage the storage."
+ )
+ parser.add_argument(
+ "--image_base_dir",
+ type = str,
+ default="",
+ help="The base directory where stored all the images."
+ )
+ parser.add_argument(
+ "--num_images",
+ type=int,
+ default=0,
+ help = "The number of images from image base dictionary"
+
+ )
+ parser.add_argument(
+ "--train_data_file",
+ type = str,
+ default=None,
+ help="The train data directory where stored all the train data,document is .json or .csv"
+ )#如果使用的是本地的数据集的话，记得一定要提供参数train_data_file和image_base_dir
+
+
+ parser.add_argument("--local_rank", type=int, default=-1, help="For distributed training: local_rank")
+ # 分布式训练时用于指定当前进程的本地编号，通常由分布式工具自动传入；单卡训练不需要关心。
+
+ if input_args is not None:
+ args = parser.parse_args(input_args)
+ else:
+ args = parser.parse_args()
+
+ if args.dataset_name is None and args.train_data_file is None:
+ raise ValueError("Specify either `--dataset_name` or `--train_data_file`")
+
+ if args.dataset_name is not None and args.train_data_file is not None:
+ raise ValueError("Specify only one of `--dataset_name` or `--train_data_file`")
+
+
+ env_local_rank = int(os.environ.get("LOCAL_RANK", -1))
+ #意思：从环境变量 LOCAL_RANK 中读取当前进程的编号。
+ #如果环境里没有这个变量，默认就是 -1。
+ if env_local_rank != -1 and env_local_rank != args.local_rank:## local_rank是在分布式训练时用于指定当前进程的本地编号，通常由分布式工具自动传入
+ args.local_rank = env_local_rank
+ # 意思：如果环境变量里读到的 LOCAL_RANK 不是 -1（说明确实是分布式训练），并且跟命令行参数 --local_rank 不一样，
+ # 那就用环境变量的 LOCAL_RANK 更新 args.local_rank，保证一致。
+ # 因为有时候命令行传的 --local_rank 可能是空的/错的，但环境变量通常是分布式工具帮你正确设好的
+
+ return args
+
+class ImageDataset(Dataset):
+
+ def __init__(
+ self,
+ image_base_dir,
+ train_data_file,
+ size=1024,
+ repeats=1,
+ center_crop=True,
+ ):
+ self.size = size
+ self.center_crop = center_crop
+
+ # if --dataset_name is provided or a metadata jsonl file is provided in the local --image_base_dir directory,
+ # we load the training data using load_dataset
+ if args.dataset_name is not None:
+ try:
+ from datasets import load_dataset
+ except ImportError:
+ raise ImportError(
+ "You are trying to load your data using the datasets library. If you wish to train using custom "
+ "captions please install the datasets library: `pip install datasets`. If you wish to load a "
+ "local folder containing images only, specify --instance_data_dir instead."
+ )
+ # Downloading and loading a dataset from the hub.
+ # See more about loading custom images at
+ # https://huggingface.co/docs/datasets/v2.0.0/en/dataset_script
+ dataset = load_dataset(
+ args.dataset_name,
+ args.dataset_config_name,
+ cache_dir=args.cache_dir,
+ )
+
+ # Preprocessing the datasets.
+ column_names = dataset["train"].column_names
+ #这段代码以及之后的处理都是把images当做PLI image数据来使用的
+
+
+ # 6. Get the column names for input/target.
+ if args.image_column is None:
+ image_column = column_names[0]
+ logger.info(f"image column defaulting to {image_column}")
+ else:
+ image_column = args.image_column
+ if image_column not in column_names:
+ raise ValueError(
+ f"`--image_column` value '{args.image_column}' not found in dataset columns. Dataset columns are: {', '.join(column_names)}"
+ )
+ images = dataset["train"][image_column]#images[i] 是一个字典或 Image 类型（取决于 dataset 本身的格式），通常已经被 datasets 自动加载为 PIL.Image 或类似对象，不是路径字符串
+
+ if args.caption_column is None:
+ caption_column = column_names[1]
+ logger.info(f"caption column defaulting to {caption_column}")
+ prompts = dataset["train"][caption_column]#这个参数是具体的prompt组成的数组
+
+ else:
+ if args.caption_column not in column_names:
+ raise ValueError(
+ f"`--caption_column` value '{args.caption_column}' not found in dataset columns. Dataset columns are: {', '.join(column_names)}"
+ )
+ prompts = dataset["train"][args.caption_column]#这个参数是具体的prompt组成的数组
+
+ else:
+ df = pd.read_excel(train_data_file)
+ column_names = df.columns.tolist()
+
+ image_column = column_names[0]
+ caption_column = column_names[1]
+ mos1_column = column_names[2]
+ mos1_pred_column = column_names[3]
+ mos2_column = column_names[4]
+ mos2_pred_column = column_names[5]
+
+ prompts = df[caption_column]#这个参数是具体的prompt组成的数组
+ mos1_pred = df[mos1_pred_column]
+ mos2_pred = df[mos2_pred_column]
+
+ self.image_base_dir = Path(image_base_dir)
+ self.train_data_file = Path(train_data_file)
+
+ if not self.image_base_dir.exists():
+ raise ValueError("Image base dir doesn't exists.")
+
+ images_path = df[image_column]
+
+ #去掉文件路径前面加的train_,test_
+ images_file = [image_path.split('_', 1)[-1] for image_path in images_path]
+
+ # #筛选出sd3.5，来计算阈值
+ # sd_score_mos1_pred = []
+ # sd_score_mos2_pred = []
+ # sd_score_mos1_all = 0
+ # sd_score_mos2_all = 0
+ # for i,image_file in enumerate(images_file):#i默认是从0开始的，也可以在enumerate中用start=某个数字来指定开始
+ # if image_file.split('/')[0] == "sd3_5_large":
+ # sd_score_mos1_pred.append(mos1_pred[i])
+ # sd_score_mos2_pred.append(mos2_pred[i])
+ # sd_score_mos1_all += float(mos1_pred[i])
+ # sd_score_mos2_all += float(mos2_pred[i])
+ # print("sd_3.5一共有",len(sd_score_mos1_pred),"张照片")
+
+ # #计算阈值
+ p=0.5
+ # thres = (p*sd_score_mos1_all+(1-p)*sd_score_mos2_all)/len(sd_score_mos1_pred)
+ # print("阈值为",thres)
+
+ #利用设置的阈值来筛选出最终进行微调的照片,并利用之前设置阈值的方式来计算最终的reward分数
+ images_file_final = []
+ prompts_final = []
+ reward_final = []
+ for i,image_file in enumerate(images_file):
+ score = p*mos1_pred[i] + (1-p)*mos2_pred[i]
+ # if score >= thres:
+ images_file_final.append(image_file)
+ prompts_final.append(prompts[i])
+ reward_final.append(score)
+
+ min_val = min(reward_final)
+ max_val = max(reward_final)
+
+ # 避免除以0的情况（所有值相同）
+ if max_val > min_val:
+ reward_final_normalized = [(x - min_val) / (max_val - min_val) for x in reward_final]
+ else:
+ reward_final_normalized = [0.0 for _ in reward_final] # 或者全设为1也可，根据需要
+ rewards = reward_final_normalized
+ reward_final = rewards
+ #images = [Image.open(os.path.join(args.image_base_dir, im_file)) for im_file in images_file]
+
+ # print("筛选之后还剩",len(images_file_final),"张照片\n")
+ # print("筛选之后还剩",len(prompts_final),"张照片\n")
+ # print("筛选之后还剩",len(reward_final),"张照片\n")
+ # 提取图片编号（如"001"），用于分组
+ def extract_prompt_index(image_path):
+ return image_path.split("/")[-1].split(".")[0] # 从 "ali_flux_schnell/001.png" 提取 "001"
+
+ # 建立编号 -> List[(image_path, reward, prompt)] 的映射
+ prompt_groups = defaultdict(list)
+ for img_path, reward, prompt in zip(images_file, reward_final, prompts_final):
+ idx = extract_prompt_index(img_path)
+ prompt_groups[idx].append((img_path, reward, prompt))
+
+ # 构造 DPO 训练数据
+
+ dpo_data = {
+ "prompt": [],
+ "win": [],
+ "loss": [],
+ "win_reward": [],
+ "loss_reward": [],
+ }
+
+ for idx, items in prompt_groups.items():
+ if len(items) < 2:
+ continue
+
+ # 打乱顺序，保证随机
+ random.shuffle(items)
+
+ # 尽可能多地构造不重复对，每次使用两张图片
+ while len(items) >= 2:
+ sample1 = items.pop()
+ sample2 = items.pop()
+
+ img1, reward1, prompt1 = sample1
+ img2, reward2, prompt2 = sample2
+
+ # 若评分相等则跳过，不放回
+ if reward1 == reward2:
+ continue
+
+ if reward1 > reward2:
+ winner, loser, prompt,win_reward,loss_reward = img1, img2, prompt1.strip(),reward1, reward2
+ else:
+ winner, loser, prompt ,win_reward,loss_reward= img2, img1, prompt2.strip(),reward2, reward1
+
+ dpo_data["prompt"].append(prompt)
+ dpo_data["win"].append(winner)
+ dpo_data["loss"].append(loser)
+ dpo_data["win_reward"].append(win_reward)
+ dpo_data["loss_reward"].append(loss_reward)
+
+ print(f"构造完成，共获得 {len(dpo_data['loss'])} 条 DPO 微调数据。")
+
+ # 假设 dpo_data 是一个包含若干字典的列表，键为 prompt / image_winner / image_loser
+ df = pd.DataFrame(dpo_data)
+
+ # 存储为 CSV 文件
+ csv_save_path = "dpo_dataset.csv"
+ df.to_csv(csv_save_path, index=False, encoding='utf-8')
+
+ print(f"DPO 微调数据已保存为 CSV 文件：{csv_save_path}")
+
+ images_win = dpo_data["win"]
+ images_loss = dpo_data["loss"]
+ prompts = dpo_data["prompt"]
+ rewards_win = dpo_data["win_reward"]
+ rewards_loss = dpo_data["loss_reward"]
+
+
+
+ self.prompts = []
+ for caption in prompts:
+ self.prompts.extend(itertools.repeat(caption, repeats))
+
+ self.images_win = []
+ for img in images_win:
+ self.images_win.extend(itertools.repeat(img, repeats))
+
+ self.images_loss = []
+ for img in images_loss:
+ self.images_loss.extend(itertools.repeat(img, repeats))
+
+ self.rewards_win = []
+ for reward in rewards_win:
+ self.rewards_win.extend(itertools.repeat(reward, repeats))
+ self.rewards_loss = []
+ for reward in rewards_loss:
+ self.rewards_loss.extend(itertools.repeat(reward, repeats))
+
+ train_resize = transforms.Resize(size, interpolation=transforms.InterpolationMode.BILINEAR)
+ train_crop = transforms.CenterCrop(size) if center_crop else transforms.RandomCrop(size)
+ train_flip = transforms.RandomHorizontalFlip(p=1.0)
+
+
+ img_transforms = []
+ img_transforms.append(train_resize)
+
+ if args.random_flip and random.random() < 0.5:
+ img_transforms.append(train_flip)
+ if args.center_crop:
+ img_transforms.append(train_crop)
+ self.image_transforms = transforms.Compose(
+ [*img_transforms, transforms.ToTensor(), transforms.Normalize([0.5], [0.5])]
+ )
+
+ self.num_images = len(self.images_win)
+ self._length = self.num_images
+ #SDXL以及它的变体中，会将一些额外的信息输入，但是SD3.5好像是没有和之前一样将这些信息输入的
+
+
+
+ # self.image_transforms = transforms.Compose(
+ # [
+ # transforms.Resize(size, interpolation=transforms.InterpolationMode.BILINEAR),
+ # transforms.CenterCrop(size) if center_crop else transforms.RandomCrop(size),
+ # transforms.ToTensor(),
+ # transforms.Normalize([0.5], [0.5]),
+ # ]
+ # )#构建图片处理函数,前面已经对数据集中的照片进行处理过了，这里是包装为了之后还需要的时候使用
+
+ def __len__(self):
+ return self._length
+#如果在类中定义了__getitem__()方法，那么他的实例对象（假设为P）就可以这样P[key]取值。当实例对象做P[key]运算时，就会调用类中的__getitem__()方法。
+#如果类把某个属性定义为序列，可以使用__getitem__()输出序列属性中的某个元素.
+ def __getitem__(self, index):
+ example = {}
+ all_images_input = []
+ images = []
+ image_win = self.images_win[index % self.num_images]
+ image_loss = self.images_loss[index % self.num_images]
+ reward_win = self.rewards_win[index % self.num_images]
+ reward_loss = self.rewards_loss[index % self.num_images]
+ images.append(image_win)
+ images.append(image_loss)
+
+ for image in images:
+ image_input = Image.open(os.path.join(args.image_base_dir, image))
+ if not image_input.mode == "RGB":
+ image_input = image_input.convert("RGB")
+ image_input = exif_transpose(image_input)#这行代码是用来矫正图片的方向的，保证图片的时序信息是正确的
+ image_input_tr = self.image_transforms(image_input)
+ all_images_input.append(image_input_tr)
+ combined_im = torch.cat(all_images_input, dim=0)
+
+ example["images_input"] = combined_im
+
+ caption = self.prompts[index % self.num_images]
+ example["prompts_input"] = caption
+ example["rewards_win"] = reward_win
+ example["rewards_loss"] = reward_loss
+
+ return example
+
+def collate_fn(examples):
+ pixel_values = [example["images_input"] for example in examples]
+ prompts = [example["prompts_input"] for example in examples]
+ rewards_win = [example["rewards_win"] for example in examples]
+ rewards_loss = [example["rewards_loss"] for example in examples]
+
+ # Concat class and instance examples for prior preservation.
+ # We do this to avoid doing two forward passes.
+
+ pixel_values = torch.stack(pixel_values)
+ pixel_values = pixel_values.to(memory_format=torch.contiguous_format).float()
+
+ batch = {"pixel_values": pixel_values, "prompts": prompts, "rewards_win": rewards_win, "rewards_loss": rewards_loss}
+ #到此已经和compbench有很多不一样的地方了，比如说我的batch里的prompt就是字符串，但是它的已经是tokenizer和pad过的了
+ return batch#其中pixel_values就是处理过的图片数据，#prompts就是提示词数据
+
+
+# class PromptDataset(Dataset):
+# "A simple dataset to prepare the prompts to generate class images on multiple GPUs."
+
+# def __init__(self, prompt, num_samples):
+# self.prompt = prompt
+# self.num_samples = num_samples
+
+# def __len__(self):
+# return self.num_samples
+
+# def __getitem__(self, index):
+# example = {}
+# example["prompt"] = self.prompt
+# example["index"] = index
+# return example
+
+
+def tokenize_prompt(tokenizer, prompt):
+ text_inputs = tokenizer(
+ prompt,
+ padding="max_length",
+ max_length=tokenizer.model_max_length,
+ truncation=True,
+ return_tensors="pt",
+ )
+ text_input_ids = text_inputs.input_ids
+ return text_input_ids
+
+
+def _encode_prompt_with_t5(
+ text_encoder,
+ tokenizer,
+ max_sequence_length,
+ prompt=None,
+ num_images_per_prompt=1,
+ device=None,
+ text_input_ids=None,
+):
+ prompt = [prompt] if isinstance(prompt, str) else prompt
+ batch_size = len(prompt)
+
+ if tokenizer is not None:
+ text_inputs = tokenizer(
+ prompt,
+ padding="max_length",
+ max_length=max_sequence_length,
+ truncation=True,
+ add_special_tokens=True,
+ return_tensors="pt",
+ )
+ text_input_ids = text_inputs.input_ids
+ else:
+ if text_input_ids is None:
+ raise ValueError("text_input_ids must be provided when the tokenizer is not specified")
+
+ prompt_embeds = text_encoder(text_input_ids.to(device))[0]
+
+ dtype = text_encoder.dtype
+
+
+
+ prompt_embeds = prompt_embeds.to(dtype=dtype, device=device)
+
+ _, seq_len, _ = prompt_embeds.shape
+
+ # duplicate text embeddings and attention mask for each generation per prompt, using mps friendly method
+ prompt_embeds = prompt_embeds.repeat(1, num_images_per_prompt, 1)
+ prompt_embeds = prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)
+
+ return prompt_embeds
+
+
+def _encode_prompt_with_clip(
+ text_encoder,
+ tokenizer,
+ prompt: str,
+ device=None,
+ text_input_ids=None,
+ num_images_per_prompt: int = 1,
+):
+ prompt = [prompt] if isinstance(prompt, str) else prompt
+ batch_size = len(prompt)
+
+ if tokenizer is not None:
+ text_inputs = tokenizer(
+ prompt,
+ padding="max_length",
+ max_length=77,
+ truncation=True,
+ return_tensors="pt",
+ )
+
+ text_input_ids = text_inputs.input_ids
+ else:
+ if text_input_ids is None:
+ raise ValueError("text_input_ids must be provided when the tokenizer is not specified")
+
+ prompt_embeds = text_encoder(text_input_ids.to(device), output_hidden_states=True)
+
+ pooled_prompt_embeds = prompt_embeds[0]
+ prompt_embeds = prompt_embeds.hidden_states[-2]
+
+ dtype = text_encoder.dtype
+
+ prompt_embeds = prompt_embeds.to(dtype=dtype, device=device)
+
+ _, seq_len, _ = prompt_embeds.shape
+ # duplicate text embeddings for each generation per prompt, using mps friendly method
+ prompt_embeds = prompt_embeds.repeat(1, num_images_per_prompt, 1)
+ prompt_embeds = prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)
+
+ return prompt_embeds, pooled_prompt_embeds
+
+
+def encode_prompt(
+ text_encoders,
+ tokenizers,
+ prompt: str,
+ max_sequence_length,
+ device=None,
+ num_images_per_prompt: int = 1,
+ text_input_ids_list=None,
+):
+ prompt = [prompt] if isinstance(prompt, str) else prompt
+
+ clip_tokenizers = tokenizers[:2]
+ clip_text_encoders = text_encoders[:2]
+
+ clip_prompt_embeds_list = []
+ clip_pooled_prompt_embeds_list = []
+ for i, (tokenizer, text_encoder) in enumerate(zip(clip_tokenizers, clip_text_encoders)):
+ prompt_embeds, pooled_prompt_embeds = _encode_prompt_with_clip(
+ text_encoder=text_encoder,
+ tokenizer=tokenizer,
+ prompt=prompt,
+ device=device if device is not None else text_encoder.device,
+ num_images_per_prompt=num_images_per_prompt,
+ text_input_ids=text_input_ids_list[i] if text_input_ids_list else None,
+ )
+ clip_prompt_embeds_list.append(prompt_embeds)
+ clip_pooled_prompt_embeds_list.append(pooled_prompt_embeds)
+
+ clip_prompt_embeds = torch.cat(clip_prompt_embeds_list, dim=-1)
+ pooled_prompt_embeds = torch.cat(clip_pooled_prompt_embeds_list, dim=-1)
+
+ t5_prompt_embed = _encode_prompt_with_t5(
+ text_encoders[-1],
+ tokenizers[-1],
+ max_sequence_length,
+ prompt=prompt,
+ num_images_per_prompt=num_images_per_prompt,
+ text_input_ids=text_input_ids_list[-1] if text_input_ids_list else None,
+ device=device if device is not None else text_encoders[-1].device,
+ )
+
+ clip_prompt_embeds = torch.nn.functional.pad(
+ clip_prompt_embeds, (0, t5_prompt_embed.shape[-1] - clip_prompt_embeds.shape[-1])
+ )
+ prompt_embeds = torch.cat([clip_prompt_embeds, t5_prompt_embed], dim=-2)
+
+ return prompt_embeds, pooled_prompt_embeds
+
+
+def main(args):
+ if args.report_to == "wandb" and args.hub_token is not None:
+ raise ValueError(
+ "You cannot use both --report_to=wandb and --hub_token due to a security risk of exposing your token."
+ " Please use `huggingface-cli login` to authenticate with the Hub."
+ )
+
+ if torch.backends.mps.is_available() and args.mixed_precision == "bf16":
+ # due to pytorch#99272, MPS does not yet support bfloat16.
+ raise ValueError(
+ "Mixed precision training with bfloat16 is not supported on MPS. Please use fp16 (recommended) or fp32 instead."
+ )
+
+ logging_dir = Path(args.output_dir, args.logging_dir)
+
+ accelerator_project_config = ProjectConfiguration(project_dir=args.output_dir, logging_dir=logging_dir)
+ kwargs = DistributedDataParallelKwargs(find_unused_parameters=True)
+ accelerator = Accelerator(
+ gradient_accumulation_steps=args.gradient_accumulation_steps,
+ mixed_precision=args.mixed_precision,
+ log_with=args.report_to,
+ project_config=accelerator_project_config,
+ kwargs_handlers=[kwargs],
+ )
+
+ # Disable AMP for MPS.
+ if torch.backends.mps.is_available():
+ accelerator.native_amp = False
+
+ if args.report_to == "wandb":
+ if not is_wandb_available():
+ raise ImportError("Make sure to install wandb if you want to use it for logging during training.")
+
+ # Make one log on every process with the configuration for debugging.
+ logging.basicConfig(
+ format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
+ datefmt="%m/%d/%Y %H:%M:%S",
+ level=logging.INFO,
+ )
+ logger.info(accelerator.state, main_process_only=False)
+ if accelerator.is_local_main_process:
+ transformers.utils.logging.set_verbosity_warning()
+ diffusers.utils.logging.set_verbosity_info()
+ else:
+ transformers.utils.logging.set_verbosity_error()
+ diffusers.utils.logging.set_verbosity_error()
+
+ # If passed along, set the training seed now.
+ if args.seed is not None:
+ set_seed(args.seed)
+
+
+ # Handle the repository creation
+ if accelerator.is_main_process:
+ if args.output_dir is not None:
+ os.makedirs(args.output_dir, exist_ok=True)
+
+ if args.push_to_hub:
+ repo_id = create_repo(
+ repo_id=args.hub_model_id or Path(args.output_dir).name,
+ exist_ok=True,
+ ).repo_id
+
+ # Load the tokenizers
+ tokenizer_one = CLIPTokenizer.from_pretrained(
+ args.pretrained_model_name_or_path,
+ subfolder="tokenizer",
+ revision=args.revision,
+ )
+ tokenizer_two = CLIPTokenizer.from_pretrained(
+ args.pretrained_model_name_or_path,
+ subfolder="tokenizer_2",
+ revision=args.revision,
+ )
+ tokenizer_three = T5TokenizerFast.from_pretrained(
+ args.pretrained_model_name_or_path,
+ subfolder="tokenizer_3",
+ revision=args.revision,
+ )
+
+ # import correct text encoder classes
+ text_encoder_cls_one = import_model_class_from_model_name_or_path(
+ args.pretrained_model_name_or_path, args.revision
+ )
+ text_encoder_cls_two = import_model_class_from_model_name_or_path(
+ args.pretrained_model_name_or_path, args.revision, subfolder="text_encoder_2"
+ )
+ text_encoder_cls_three = import_model_class_from_model_name_or_path(
+ args.pretrained_model_name_or_path, args.revision, subfolder="text_encoder_3"
+ )
+
+ # Load scheduler and models
+ noise_scheduler = FlowMatchEulerDiscreteScheduler.from_pretrained(
+ args.pretrained_model_name_or_path, subfolder="scheduler"
+ )
+ noise_scheduler_copy = copy.deepcopy(noise_scheduler)
+ text_encoder_one, text_encoder_two, text_encoder_three = load_text_encoders(
+ text_encoder_cls_one, text_encoder_cls_two, text_encoder_cls_three
+ )
+ vae = AutoencoderKL.from_pretrained(
+ args.pretrained_model_name_or_path,
+ subfolder="vae",
+ revision=args.revision,
+ variant=args.variant,
+ )
+ transformer = SD3Transformer2DModel.from_pretrained(
+ args.pretrained_model_name_or_path, subfolder="transformer", revision=args.revision, variant=args.variant
+ )
+
+ transformer.requires_grad_(False)
+ vae.requires_grad_(False)
+ text_encoder_one.requires_grad_(False)
+ text_encoder_two.requires_grad_(False)
+ text_encoder_three.requires_grad_(False)
+
+ # For mixed precision training we cast all non-trainable weights (vae, non-lora text_encoder and non-lora transformer) to half-precision
+ # as these weights are only used for inference, keeping weights in full precision is not required.
+ weight_dtype = torch.float32
+ if accelerator.mixed_precision == "fp16":
+ weight_dtype = torch.float16
+ elif accelerator.mixed_precision == "bf16":
+ weight_dtype = torch.bfloat16
+
+ if torch.backends.mps.is_available() and weight_dtype == torch.bfloat16:
+ # due to pytorch#99272, MPS does not yet support bfloat16.
+ raise ValueError(
+ "Mixed precision training with bfloat16 is not supported on MPS. Please use fp16 (recommended) or fp32 instead."
+ )
+
+ vae.to(accelerator.device, dtype=torch.float32)
+ transformer.to(accelerator.device, dtype=weight_dtype)
+ text_encoder_one.to(accelerator.device, dtype=weight_dtype)
+ text_encoder_two.to(accelerator.device, dtype=weight_dtype)
+ text_encoder_three.to(accelerator.device, dtype=weight_dtype)
+
+ if args.gradient_checkpointing:
+ transformer.enable_gradient_checkpointing()
+ if args.train_text_encoder:
+ text_encoder_one.gradient_checkpointing_enable()
+ text_encoder_two.gradient_checkpointing_enable()
+ if args.lora_layers is not None:
+ target_modules = [layer.strip() for layer in args.lora_layers.split(",")]
+ else:
+ target_modules = [
+ "attn.add_k_proj",
+ "attn.add_q_proj",
+ "attn.add_v_proj",
+ "attn.to_add_out",
+ "attn.to_k",
+ "attn.to_out.0",
+ "attn.to_q",
+ "attn.to_v",
+ ]
+ if args.lora_blocks is not None:
+ target_blocks = [int(block.strip()) for block in args.lora_blocks.split(",")]
+ target_modules = [
+ f"transformer_blocks.{block}.{module}" for block in target_blocks for module in target_modules
+ ]
+
+ # now we will add new LoRA weights to the attention layers
+ transformer_lora_config = LoraConfig(
+ r=args.rank,
+ lora_alpha=args.rank,
+ init_lora_weights="gaussian",
+ target_modules=target_modules,
+ )
+ transformer.add_adapter(transformer_lora_config)
+
+ if args.train_text_encoder:
+ text_lora_config = LoraConfig(
+ r=args.rank,
+ lora_alpha=args.rank,
+ init_lora_weights="gaussian",
+ target_modules=["q_proj", "k_proj", "v_proj", "out_proj"],
+ )
+ text_encoder_one.add_adapter(text_lora_config)
+ text_encoder_two.add_adapter(text_lora_config)
+
+ def unwrap_model(model):
+ model = accelerator.unwrap_model(model)
+ model = model._orig_mod if is_compiled_module(model) else model
+ return model
+
+ # create custom saving & loading hooks so that `accelerator.save_state(...)` serializes in a nice format
+ def save_model_hook(models, weights, output_dir):
+ if accelerator.is_main_process:
+ transformer_lora_layers_to_save = None
+ text_encoder_one_lora_layers_to_save = None
+ text_encoder_two_lora_layers_to_save = None
+
+ for model in models:
+ if isinstance(unwrap_model(model), type(unwrap_model(transformer))):
+ model = unwrap_model(model)
+ if args.upcast_before_saving:
+ model = model.to(torch.float32)
+ transformer_lora_layers_to_save = get_peft_model_state_dict(model)
+ elif args.train_text_encoder and isinstance(
+ unwrap_model(model), type(unwrap_model(text_encoder_one))
+ ): # or text_encoder_two
+ # both text encoders are of the same class, so we check hidden size to distinguish between the two
+ model = unwrap_model(model)
+ hidden_size = model.config.hidden_size
+ if hidden_size == 768:
+ text_encoder_one_lora_layers_to_save = get_peft_model_state_dict(model)
+ elif hidden_size == 1280:
+ text_encoder_two_lora_layers_to_save = get_peft_model_state_dict(model)
+ else:
+ raise ValueError(f"unexpected save model: {model.__class__}")
+
+ # make sure to pop weight so that corresponding model is not saved again
+ if weights:
+ weights.pop()
+
+ StableDiffusion3Pipeline.save_lora_weights(
+ output_dir,
+ transformer_lora_layers=transformer_lora_layers_to_save,
+ text_encoder_lora_layers=text_encoder_one_lora_layers_to_save,
+ text_encoder_2_lora_layers=text_encoder_two_lora_layers_to_save,
+ )
+ #加载模型函数load_model_hook其实就是专门为从之前的断点中加载信息而设置的，在别的地方不会被调用，因为他加载的只是微调之后的一些参数，并不是全部的参数，模型本身还是会从一开始的MODEL_NAME="/home/wangjiarui/zy/sd3_5/checkpoints"
+ #中加载
+ #当您不使用 --resume_from_checkpoint 时，accelerator 不会使用load_model_hook加载检查点，而是直接使用 accelerator.prepare 准备从预训练模型开始训练。
+ # 调用条件：因此，load_model_hook 只有在以下情况下被调用：
+ # 训练过程使用了 accelerator.load_state。
+ # 这是通过 --resume_from_checkpoint 参数触发的，通常表示用户希望从之前的训练断点继续训练，而不是从头开始。
+ def load_model_hook(models, input_dir):
+ transformer_ = None
+ text_encoder_one_ = None
+ text_encoder_two_ = None
+
+ if not accelerator.distributed_type == DistributedType.DEEPSPEED:
+ print("不是DistributedType.DEEPSPEED！！！！！！！！！！！！")
+ while len(models) > 0:
+ model = models.pop()
+
+ if isinstance(unwrap_model(model), type(unwrap_model(transformer))):
+ transformer_ = unwrap_model(model)
+ elif isinstance(unwrap_model(model), type(unwrap_model(text_encoder_one))):
+ text_encoder_one_ = unwrap_model(model)
+ elif isinstance(unwrap_model(model), type(unwrap_model(text_encoder_two))):
+ text_encoder_two_ = unwrap_model(model)
+ else:
+ raise ValueError(f"unexpected save model: {model.__class__}")
+ print([type(unwrap_model(m)) for m in models])
+
+
+ else:
+ transformer_ = SD3Transformer2DModel.from_pretrained(
+ args.pretrained_model_name_or_path, subfolder="transformer"
+ )
+ transformer_.add_adapter(transformer_lora_config)
+ if args.train_text_encoder:
+ text_encoder_one_ = text_encoder_cls_one.from_pretrained(
+ args.pretrained_model_name_or_path, subfolder="text_encoder"
+ )
+ text_encoder_two_ = text_encoder_cls_two.from_pretrained(
+ args.pretrained_model_name_or_path, subfolder="text_encoder_2"
+ )
+
+ lora_state_dict = StableDiffusion3Pipeline.lora_state_dict(input_dir)
+
+ lora_state_dict = StableDiffusion3Pipeline.lora_state_dict(input_dir)
+
+ transformer_state_dict = {
+ f'{k.replace("transformer.", "")}': v for k, v in lora_state_dict.items() if k.startswith("transformer.")
+ }
+ transformer_state_dict = convert_unet_state_dict_to_peft(transformer_state_dict)
+ incompatible_keys = set_peft_model_state_dict(transformer_, transformer_state_dict, adapter_name="default")
+ if incompatible_keys is not None:
+ # check only for unexpected keys
+ unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None)
+ if unexpected_keys:
+ logger.warning(
+ f"Loading adapter weights from state_dict led to unexpected keys not found in the model: "
+ f" {unexpected_keys}. "
+ )
+ if args.train_text_encoder:
+ # Do we need to call `scale_lora_layers()` here?
+ _set_state_dict_into_text_encoder(lora_state_dict, prefix="text_encoder.", text_encoder=text_encoder_one_)
+
+ _set_state_dict_into_text_encoder(
+ lora_state_dict, prefix="text_encoder_2.", text_encoder=text_encoder_two_
+ )
+
+ # Make sure the trainable params are in float32. This is again needed since the base models
+ # are in `weight_dtype`. More details:
+ # https://github.com/huggingface/diffusers/pull/6514#discussion_r1449796804
+ if args.mixed_precision == "fp16":
+ models = [transformer_]
+ if args.train_text_encoder:
+ models.extend([text_encoder_one_, text_encoder_two_])
+ # only upcast trainable parameters (LoRA) into fp32
+ cast_training_params(models)
+
+ accelerator.register_save_state_pre_hook(save_model_hook)
+ accelerator.register_load_state_pre_hook(load_model_hook)
+
+ # Enable TF32 for faster training on Ampere GPUs,
+ # cf https://pytorch.org/docs/stable/notes/cuda.html#tensorfloat-32-tf32-on-ampere-devices
+ if args.allow_tf32 and torch.cuda.is_available():
+ torch.backends.cuda.matmul.allow_tf32 = True
+
+ if args.scale_lr:
+ args.learning_rate = (
+ args.learning_rate * args.gradient_accumulation_steps * args.train_batch_size * accelerator.num_processes
+ )
+
+ # Make sure the trainable params are in float32.
+ if args.mixed_precision == "fp16":
+ models = [transformer]
+ if args.train_text_encoder:
+ models.extend([text_encoder_one, text_encoder_two])
+ # only upcast trainable parameters (LoRA) into fp32
+ cast_training_params(models, dtype=torch.float32)
+
+ transformer_lora_parameters = list(filter(lambda p: p.requires_grad, transformer.parameters()))
+ if args.train_text_encoder:
+ text_lora_parameters_one = list(filter(lambda p: p.requires_grad, text_encoder_one.parameters()))
+ text_lora_parameters_two = list(filter(lambda p: p.requires_grad, text_encoder_two.parameters()))
+
+ # Optimization parameters
+ transformer_parameters_with_lr = {"params": transformer_lora_parameters, "lr": args.learning_rate}
+ if args.train_text_encoder:
+ # different learning rate for text encoder and unet
+ text_lora_parameters_one_with_lr = {
+ "params": text_lora_parameters_one,
+ "weight_decay": args.adam_weight_decay_text_encoder,
+ "lr": args.text_encoder_lr if args.text_encoder_lr else args.learning_rate,
+ }
+ text_lora_parameters_two_with_lr = {
+ "params": text_lora_parameters_two,
+ "weight_decay": args.adam_weight_decay_text_encoder,
+ "lr": args.text_encoder_lr if args.text_encoder_lr else args.learning_rate,
+ }
+ params_to_optimize = [
+ transformer_parameters_with_lr,
+ text_lora_parameters_one_with_lr,
+ text_lora_parameters_two_with_lr,
+ ]
+ else:
+ params_to_optimize = [transformer_parameters_with_lr]
+
+ # Optimizer creation
+ if not (args.optimizer.lower() == "prodigy" or args.optimizer.lower() == "adamw"):
+ logger.warning(
+ f"Unsupported choice of optimizer: {args.optimizer}.Supported optimizers include [adamW, prodigy]."
+ "Defaulting to adamW"
+ )
+ args.optimizer = "adamw"
+
+ if args.use_8bit_adam and not args.optimizer.lower() == "adamw":
+ logger.warning(
+ f"use_8bit_adam is ignored when optimizer is not set to 'AdamW'. Optimizer was "
+ f"set to {args.optimizer.lower()}"
+ )
+
+ if args.optimizer.lower() == "adamw":
+ if args.use_8bit_adam:
+ try:
+ import bitsandbytes as bnb
+ except ImportError:
+ raise ImportError(
+ "To use 8-bit Adam, please install the bitsandbytes library: `pip install bitsandbytes`."
+ )
+
+ optimizer_class = bnb.optim.AdamW8bit
+ else:
+ optimizer_class = torch.optim.AdamW
+
+ optimizer = optimizer_class(
+ params_to_optimize,
+ betas=(args.adam_beta1, args.adam_beta2),
+ weight_decay=args.adam_weight_decay,
+ eps=args.adam_epsilon,
+ )
+
+ if args.optimizer.lower() == "prodigy":
+ try:
+ import prodigyopt
+ except ImportError:
+ raise ImportError("To use Prodigy, please install the prodigyopt library: `pip install prodigyopt`")
+
+ optimizer_class = prodigyopt.Prodigy
+
+ if args.learning_rate <= 0.1:
+ logger.warning(
+ "Learning rate is too low. When using prodigy, it's generally better to set learning rate around 1.0"
+ )
+ if args.train_text_encoder and args.text_encoder_lr:
+ logger.warning(
+ f"Learning rates were provided both for the transformer and the text encoder- e.g. text_encoder_lr:"
+ f" {args.text_encoder_lr} and learning_rate: {args.learning_rate}. "
+ f"When using prodigy only learning_rate is used as the initial learning rate."
+ )
+ # changes the learning rate of text_encoder_parameters_one and text_encoder_parameters_two to be
+ # --learning_rate
+ params_to_optimize[1]["lr"] = args.learning_rate
+ params_to_optimize[2]["lr"] = args.learning_rate
+
+ optimizer = optimizer_class(
+ params_to_optimize,
+ betas=(args.adam_beta1, args.adam_beta2),
+ beta3=args.prodigy_beta3,
+ weight_decay=args.adam_weight_decay,
+ eps=args.adam_epsilon,
+ decouple=args.prodigy_decouple,
+ use_bias_correction=args.prodigy_use_bias_correction,
+ safeguard_warmup=args.prodigy_safeguard_warmup,
+ )
+
+
+ # Dataset and DataLoaders creation:
+ train_dataset = ImageDataset(
+ image_base_dir=args.image_base_dir,
+ train_data_file=args.train_data_file,
+ size=args.resolution,
+ repeats=args.repeats,
+ center_crop=args.center_crop,
+ )
+ #每个进程独立运行 train_gors_lora_sd3.py，执行相同的代码，
+ #ImageDataset 是在每个进程的 Python 环境中创建的，accelerate 不会自动共享 train_dataset 的内存。
+ #因此一开始imagedataset类不能够装载那些占据内存量很大的数据，只能保存一些元数据，不然就会占用特别多的内存
+ #之前dreambooth由于只需要几张照片就能实现全局微调，所以将那几张照片在imagedataset类中就放到了cpu中，并没有像别的训练代码一样在真正处理batch的时候才加载
+
+
+ train_dataloader = torch.utils.data.DataLoader(
+ train_dataset,
+ batch_size=args.train_batch_size,
+ shuffle=True,
+ collate_fn=lambda examples: collate_fn(examples),#样本（examples）是 DataLoader 从 Dataset 中一个一个读取出来的。根据 batch_size 拿出一批样本
+ num_workers=args.dataloader_num_workers,
+ )#用torch自带的DataLoader函数对dataset进行切割
+ #这里只是设置了个数据加载器，相当于是初始化，并没有真正的用这个东西来处理数据，这个东西是后面训练的时候对train_dataloader进行遍历的时候
+ #会调用__getitem__()函数来对train_dataset进行train_dataset[index]进行索引取出examples，然后用collate_fn进行处理，得到batch
+ #实际调用的代码是：for step, batch in enumerate(train_dataloader)
+
+ if not args.train_text_encoder:
+ tokenizers = [tokenizer_one, tokenizer_two, tokenizer_three]
+ text_encoders = [text_encoder_one, text_encoder_two, text_encoder_three]
+
+ def compute_text_embeddings(prompt, text_encoders, tokenizers):
+ with torch.no_grad():
+ prompt_embeds, pooled_prompt_embeds = encode_prompt(
+ text_encoders, tokenizers, prompt, args.max_sequence_length
+ )
+ prompt_embeds = prompt_embeds.to(accelerator.device)
+ #每个 token 对应的向量，形状类似于 [batch, seq_len, dim]
+ pooled_prompt_embeds = pooled_prompt_embeds.to(accelerator.device)
+ #整个句子压缩成一个向量，通常用于条件控制（形状 [batch, dim]）
+ return prompt_embeds, pooled_prompt_embeds
+
+
+ vae_config_shift_factor = vae.config.shift_factor
+ vae_config_scaling_factor = vae.config.scaling_factor
+
+
+ # Scheduler and math around the number of training steps.
+ overrode_max_train_steps = False
+ num_update_steps_per_epoch = math.ceil(len(train_dataloader) / args.gradient_accumulation_steps)
+ if args.max_train_steps is None:
+ args.max_train_steps = args.num_train_epochs * num_update_steps_per_epoch
+ overrode_max_train_steps = True
+
+ lr_scheduler = get_scheduler(
+ args.lr_scheduler,
+ optimizer=optimizer,
+ num_warmup_steps=args.lr_warmup_steps * accelerator.num_processes,
+ num_training_steps=args.max_train_steps * accelerator.num_processes,
+ num_cycles=args.lr_num_cycles,
+ power=args.lr_power,
+ )
+
+ # Prepare everything with our `accelerator`.
+ if args.train_text_encoder:
+ (
+ transformer,
+ text_encoder_one,
+ text_encoder_two,
+ optimizer,
+ train_dataloader,
+ lr_scheduler,
+ ) = accelerator.prepare(
+ transformer, text_encoder_one, text_encoder_two, optimizer, train_dataloader, lr_scheduler
+ )#这里对text_encoder_one, text_encoder_two进行了包装，所以之后就不能够直接访问模型的类型了
+ assert text_encoder_one is not None
+ assert text_encoder_two is not None
+ assert text_encoder_three is not None
+ else:
+ transformer, optimizer, train_dataloader, lr_scheduler = accelerator.prepare(
+ transformer, optimizer, train_dataloader, lr_scheduler
+ )
+ # if args.cache_latents:
+ # if (os.path.exists("./latents_cache/latent_1.pt")):
+ # if args.cache_latents:
+ # latents_cache = [
+ # torch.load(f"./latents_cache/latent_{i}.pt").to(accelerator.device, dtype=weight_dtype)
+ # for i in range(len(train_dataset))
+ # ]
+ # else:
+ # latents_cache = []
+ # for batch_idx,batch in enumerate(tqdm(train_dataloader, desc="Caching latents")):
+ # # if batch_idx <=3198:
+ # # continue
+ # with torch.no_grad():
+ # batch["pixel_values"] = batch["pixel_values"].to(
+ # accelerator.device, non_blocking=True, dtype=vae.dtype
+ # )
+ # latent = vae.encode(batch["pixel_values"]).latent_dist
+ # latents_cache.append(latent)
+ # torch.save(latent.sample().cpu(), f"./latents_cache/latent_{batch_idx}.pt")
+ # # 保存潜变量
+ # #可是这样就相当于是在一个GPU一次放了所有的照片，运行了之后发现会GPU显存不够用
+
+ # if args.validation_prompts is None:
+ # del vae
+ # free_memory()
+
+ # We need to recalculate our total training steps as the size of the training dataloader may have changed.
+ num_update_steps_per_epoch = math.ceil(len(train_dataloader) / args.gradient_accumulation_steps)
+ print(f"每周期训练步骤是{num_update_steps_per_epoch}!!!!!!!!")
+ if overrode_max_train_steps:
+ args.max_train_steps = args.num_train_epochs * num_update_steps_per_epoch
+ # Afterwards we recalculate our number of training epochs
+ args.num_train_epochs = math.ceil(args.max_train_steps / num_update_steps_per_epoch)
+
+ # We need to initialize the trackers we use, and also store our configuration.
+ # The trackers initializes automatically on the main process.
+ if accelerator.is_main_process:
+ tracker_name = "gors-sd3-lora"
+ accelerator.init_trackers(tracker_name, config=vars(args))
+
+ # Train!
+ total_batch_size = args.train_batch_size * accelerator.num_processes * args.gradient_accumulation_steps
+#----------------------------------------------------------------------------
+ logger.info("***** Running training *****")
+ logger.info(f" Num examples = {len(train_dataset)}")
+ logger.info(f" Num batches each epoch = {len(train_dataloader)}")
+ logger.info(f" Num Epochs = {args.num_train_epochs}")
+ logger.info(f" Instantaneous batch size per device = {args.train_batch_size}")
+ logger.info(f" Total train batch size (w. parallel, distributed & accumulation) = {total_batch_size}")
+ logger.info(f" Gradient Accumulation steps = {args.gradient_accumulation_steps}")
+ logger.info(f" Total optimization steps = {args.max_train_steps}")
+ global_step = 0
+ first_epoch = 0
+ #这一部分代码应该也需要相应的更改一下
+#-----------------------------------------------------------------------------------
+
+ # Potentially load in the weights and states from a previous save
+ if args.resume_from_checkpoint:
+ if args.resume_from_checkpoint != "latest":
+ path = os.path.basename(args.resume_from_checkpoint)
+ else:
+ # Get the mos recent checkpoint
+ dirs = os.listdir(args.output_dir)
+ dirs = [d for d in dirs if d.startswith("checkpoint")]
+ dirs = sorted(dirs, key=lambda x: int(x.split("-")[1]))
+ path = dirs[-1] if len(dirs) > 0 else None
+
+ if path is None:
+ accelerator.print(
+ f"Checkpoint '{args.resume_from_checkpoint}' does not exist. Starting a new training run."
+ )
+ args.resume_from_checkpoint = None
+ initial_global_step = 0
+ else:
+ accelerator.print(f"Resuming from checkpoint {path}")
+ accelerator.load_state(os.path.join(args.output_dir, path), model_hook=load_model_hook)
+ #加载模型其实就是专门为从之前的断点中加载信息而设置的，在别的地方不会被调用，因为他加载的只是微调之后的一些参数，并不是全部的参数，模型本身还是会从一开始的MODEL_NAME="/home/wangjiarui/zy/sd3_5/checkpoints"
+ #中加载
+ #当您不使用 --resume_from_checkpoint 时，accelerator 不会加载检查点，而是直接使用 accelerator.prepare 准备从预训练模型开始训练。
+ global_step = int(path.split("-")[1])
+
+ initial_global_step = global_step
+ resume_global_step = global_step * args.gradient_accumulation_steps
+ first_epoch = global_step // num_update_steps_per_epoch
+ resume_step = resume_global_step % (num_update_steps_per_epoch * args.gradient_accumulation_steps)
+ #断点
+
+ else:
+ initial_global_step = 0
+
+ progress_bar = tqdm(
+ range(0, args.max_train_steps),
+ initial=initial_global_step,
+ desc="Steps",
+ # Only show the progress bar once on each machine.
+ disable=not accelerator.is_local_main_process,
+ )
+
+ def get_sigmas(timesteps, n_dim=4, dtype=torch.float32):
+ sigmas = noise_scheduler_copy.sigmas.to(device=accelerator.device, dtype=dtype)
+ schedule_timesteps = noise_scheduler_copy.timesteps.to(accelerator.device)
+ timesteps = timesteps.to(accelerator.device)
+ step_indices = [(schedule_timesteps == t).nonzero().item() for t in timesteps]
+
+ sigma = sigmas[step_indices].flatten()
+ while len(sigma.shape) < n_dim:
+ sigma = sigma.unsqueeze(-1)
+ return sigma
+#------------------------------------------------------------------
+#整个训练过程的主体
+ for epoch in range(first_epoch, args.num_train_epochs):
+ #--------------------------------------------------------------
+ #每个周期
+ transformer.train()
+ if args.train_text_encoder:
+ text_encoder_one.train()
+ text_encoder_two.train()
+
+ # set top parameter requires_grad = True for gradient checkpointing works
+ accelerator.unwrap_model(text_encoder_one).text_model.embeddings.requires_grad_(True)
+ accelerator.unwrap_model(text_encoder_two).text_model.embeddings.requires_grad_(True)
+
+ for step, batch in enumerate(train_dataloader):
+ #-----------------------------------------------------------
+ #每个周期里的一次batch
+
+ # Skip steps until we reach the resumed step
+ if args.resume_from_checkpoint and epoch == first_epoch and step < resume_step:
+ continue
+ models_to_accumulate = [transformer]
+ if args.train_text_encoder:
+ models_to_accumulate.extend([text_encoder_one, text_encoder_two])
+ with accelerator.accumulate(models_to_accumulate):
+ prompts = batch["prompts"]
+
+ # encode batch prompts when custom prompts are provided for each image -
+
+ if not args.train_text_encoder:
+ prompt_embeds, pooled_prompt_embeds = compute_text_embeddings(
+ prompts, text_encoders, tokenizers
+ )
+ else:
+ tokens_one = tokenize_prompt(tokenizer_one, prompts)
+ tokens_two = tokenize_prompt(tokenizer_two, prompts)
+ tokens_three = tokenize_prompt(tokenizer_three, prompts)
+ text_encoders=[accelerator.unwrap_model(text_encoder_one), accelerator.unwrap_model(text_encoder_two), text_encoder_three]
+ #text_encoders_upwrap = [accelerator.unwrap_model(text_encoder) for text_encoder in text_encoders]
+
+ prompt_embeds, pooled_prompt_embeds = encode_prompt(
+ text_encoders,
+ tokenizers=[None, None, None],
+ prompt=prompts,
+ max_sequence_length=args.max_sequence_length,
+ text_input_ids_list=[tokens_one, tokens_two, tokens_three],
+ )#由于compute_text_embeddings是专门在 not args.train_text_encoder情况下设置的函数，是无梯度运算
+ #所以在args.train_text_encoder的情况下还需要再边写一遍差不多的逻辑，但是是梯度运算
+
+
+ # Convert images to latent space
+ # if args.cache_latents:
+ # model_input = latents_cache[step].sample()
+ # else:
+ pixel_values = batch["pixel_values"].to(dtype=vae.dtype)
+
+ #这个数组的维度是(batch_size, 2 * C, H, W)
+ #因为拼接了win和loss图片
+
+ feed_pixel_values = torch.cat(pixel_values.chunk(2, dim=1))
+ #这个操作的意思是把 pixel_values 沿着 dim=1（通道维度） 分成 2 块，然后再【拼起来
+ #.chunk(n, dim) 返回一个列表，长度为 n
+ #torch.cat([])是将这两个 tensor 沿 batch 维度 dim=0 拼接起来
+ #最终输出的维度是feed_pixel_values.shape == (2*B, C, H, W)
+ model_input = []
+ #i是从0到2*batch_size，步长是args.vae_encode_batch_size，也就是batch_size
+ for i in range(0, feed_pixel_values.shape[0], args.train_batch_size):
+ model_input.append(
+ vae.encode(feed_pixel_values[i : i + args.train_batch_size]).latent_dist.sample()
+ )
+ model_input = torch.cat(model_input, dim=0)
+ model_input = (model_input- vae_config_shift_factor) * vae.config.scaling_factor
+
+ model_input = model_input.to(dtype=weight_dtype)
+ #SD3 中对 latent 有归一化处理，shift 和 scale 是根据 VAE 配置做的。
+ #最后转换成 weight_dtype，通常是 float16 或 bfloat16。
+
+ # Sample noise that we'll add to the model_input
+ # Sample noise that we'll add to the model_input
+ noise = torch.randn_like(model_input).chunk(2)[0].repeat(2, 1, 1, 1)
+ bsz = model_input.shape[0] // 2
+ #为每个 latent 样本生成随机噪声 z1，用于构造 noisy latent zt。
+ #bsz 是 batch size。
+
+ # Sample a random timestep for each image
+ # for weighting schemes where we sample timesteps non-uniformly
+ u = compute_density_for_timestep_sampling(
+ weighting_scheme=args.weighting_scheme,
+ batch_size=bsz,
+ logit_mean=args.logit_mean,
+ logit_std=args.logit_std,
+ mode_scale=args.mode_scale,
+ )
+ indices = (u * noise_scheduler_copy.config.num_train_timesteps).long()
+ timesteps = noise_scheduler_copy.timesteps[indices].to(device=model_input.device).repeat(2)
+
+ # 为每张图像采样一个时间步 t，决定该步噪声水平。
+ # 不同的 weighting_scheme 决定 u 的分布，可能是均匀、高斯、偏置等。
+ # 得到 timesteps，用于 transformer 推理时的条件。
+
+ # Add noise according to flow matching.
+ # zt = (1 - texp) * x + texp * z1
+ sigmas = get_sigmas(timesteps, n_dim=model_input.ndim, dtype=model_input.dtype)
+ noisy_model_input = (1.0 - sigmas) * model_input + sigmas * noise
+ # 这是 flow matching 的核心公式：
+ # z_t = (1 - σ) * x + σ * z1
+ # 其中 x 是干净 latent，z1 是噪声，σ 是对应时间步的噪声比例（可能近似 exp(t)）。
+ # 这与普通 diffusion 中 x_t = √α x₀ + √(1-α) ε 不同，是一种更简单、更稳定的插值结构。
+ prompt_embeds = prompt_embeds.repeat(2, 1, 1)
+ pooled_prompt_embeds = pooled_prompt_embeds.repeat(2, 1)
+
+
+ # Predict the noise residual
+ #就相当于，在一个干净的图片上加上随机噪声，然后让模型进行预测
+ #这里的预测分为两种，一种是预测噪声，一种是预测原本的图片
+ model_pred = transformer(
+ hidden_states=noisy_model_input,
+ timestep=timesteps,
+ encoder_hidden_states=prompt_embeds,
+ pooled_projections=pooled_prompt_embeds,
+ return_dict=False,
+ )[0]
+ # 将 noisy latent z_t 输入 transformer。
+ # 模型要学的是：根据 prompt，推断出当前 t 对应的 目标向量（通常是噪声或 clean latent）。
+ rewards_win = torch.FloatTensor(batch["rewards_win"]).squeeze().to(model_pred.device)
+ rewards_loss = torch.FloatTensor(batch["rewards_loss"]).squeeze().to(model_pred.device)
+
+ # Follow: Section 5 of https://arxiv.org/abs/2206.00364.
+ # Preconditioning of the model outputs.
+ if args.precondition_outputs:
+ model_pred = model_pred * (-sigmas) + noisy_model_input
+ #这一步相当于，如果unet是预测噪声的，那这里就去掉噪声
+ #不然就是预测原本图像的，那就不用处理
+
+ # these weighting schemes use a uniform timestep sampling
+ # and instead post-weight the loss
+ weighting = compute_loss_weighting_for_sd3(weighting_scheme=args.weighting_scheme, sigmas=sigmas)
+
+ # flow matching loss
+ if args.precondition_outputs:
+ target = model_input
+ else:
+ target = noise - model_input
+ # target 是 ground truth，根据是否预处理输出有不同：
+ # 如果用了 preconditioning，目标是 clean latent x
+ # 否则是 z1 - x，从 flow matching 推导而来
+ # weighting 是 timestep 对损失的权重函数。与 sigmas 有关，不同 scheme 下权重不同。
+
+
+ # Compute regular loss.
+ per_pixel_loss = (weighting.float() * (model_pred.float() - target.float()) ** 2)
+
+ model_losses = per_pixel_loss.mean(dim=list(range(1, len(per_pixel_loss.shape)))) # [2B]
+ model_losses_w, model_losses_l = model_losses.chunk(2)
+ raw_model_loss = 0.5 * (model_losses_w.mean() + model_losses_l.mean())
+ model_diff = model_losses_w - model_losses_l # These are both LBS (as is t)
+ # Reference model predictions.
+ accelerator.unwrap_model(transformer).disable_adapters()
+ with torch.no_grad():
+ ref_pred = transformer(
+ hidden_states=noisy_model_input,
+ timestep=timesteps,
+ encoder_hidden_states=prompt_embeds,
+ pooled_projections=pooled_prompt_embeds,
+ return_dict=False,
+ )[0].detach()#.detach() 返回的是一个“不需要梯度”的新张量，它共享原始数据，但不会再被自动求导追踪。
+
+ if args.precondition_outputs:
+ ref_pred = ref_pred * (-sigmas) + noisy_model_input
+ # Compute regular loss.
+ ref_per_pixel_loss = (weighting.float() * (ref_pred.float() - target.float()) ** 2)
+
+ ref_losses = ref_per_pixel_loss.mean(dim=list(range(1, len(ref_per_pixel_loss.shape)))) # [2B]
+ ref_losses_w, ref_losses_l = ref_losses.chunk(2)
+ raw_ref_loss = 0.5 * (ref_losses_w.mean() + ref_losses_l.mean())
+ ref_diff = ref_losses_w - ref_losses_l # These are both LBS (as is t)
+ # Reference model predictions.
+
+ # Re-enable adapters.
+ accelerator.unwrap_model(transformer).enable_adapters()
+
+ # Final loss.
+ # logits = ref_diff - model_diff
+ # logits = ref_losses_w - ref_losses_l - (model_losses_w - model_losses_l)
+ first_loss = (model_losses_w - ref_losses_w)*torch.exp(rewards_win)
+ second_loss = (model_losses_l - ref_losses_l)*torch.exp(rewards_loss)
+ logits = second_loss - first_loss
+
+ loss = -1 * F.logsigmoid(0.5*args.beta_dpo * logits).mean()
+
+
+ accelerator.backward(loss)
+
+
+ #反向传播，对梯度进行更新
+ if accelerator.sync_gradients:
+ params_to_clip = (
+ itertools.chain(
+ transformer_lora_parameters, text_lora_parameters_one, text_lora_parameters_two
+ )
+ if args.train_text_encoder
+ else transformer_lora_parameters
+ )
+ accelerator.clip_grad_norm_(params_to_clip, args.max_grad_norm)
+
+ optimizer.step()
+ lr_scheduler.step()
+ optimizer.zero_grad()
+
+
+
+ #这是在 分布式训练（尤其是多卡训练） 中常见的一种写法，用来确保 只有在真正执行了梯度同步（即一次有效的优化步骤）时 才更新进度条和 global_step
+ # 🔹 accelerator.sync_gradients
+ # 这是 HuggingFace Accelerate 提供的属性，表示当前是否是 同步梯度的步骤：
+ # 如果你使用 梯度累积（gradient accumulation），那么并不是每个 batch 都会立刻做一次 optimizer.step()。
+ # 只有当累积步数达到设定值（如 accumulation_steps = 4），才会真正进行一次反向传播 + 梯度同步 + 参数更新。
+ #batch是比global_step更小的步骤，global_step是每次梯度更新会加1，batch是每次处理一批数据就是一个batch
+
+ # Checks if the accelerator has performed an optimization step behind the scenes
+ if accelerator.sync_gradients:
+ progress_bar.update(1)
+ global_step += 1
+
+ if accelerator.is_main_process or accelerator.distributed_type == DistributedType.DEEPSPEED:
+ #看看是否要存储当前的参数
+ if global_step % args.checkpointing_steps == 0:
+ # _before_ saving state, check if this save would set us over the `checkpoints_total_limit`
+ if args.checkpoints_total_limit is not None:
+ checkpoints = os.listdir(args.output_dir)
+ checkpoints = [d for d in checkpoints if d.startswith("checkpoint")]
+ checkpoints = sorted(checkpoints, key=lambda x: int(x.split("-")[1]))
+
+ # before we save the new checkpoint, we need to have at _most_ `checkpoints_total_limit - 1` checkpoints
+ if len(checkpoints) >= args.checkpoints_total_limit:
+ num_to_remove = len(checkpoints) - args.checkpoints_total_limit + 1
+ removing_checkpoints = checkpoints[0:num_to_remove]
+
+ logger.info(
+ f"{len(checkpoints)} checkpoints already exist, removing {len(removing_checkpoints)} checkpoints"
+ )
+ logger.info(f"removing checkpoints: {', '.join(removing_checkpoints)}")
+
+ for removing_checkpoint in removing_checkpoints:
+ removing_checkpoint = os.path.join(args.output_dir, removing_checkpoint)
+ shutil.rmtree(removing_checkpoint)
+
+ save_path = os.path.join(args.output_dir, f"checkpoint-{global_step}")
+ accelerator.save_state(save_path)
+ logger.info(f"Saved state to {save_path}")
+ if accelerator.is_main_process:
+ if args.validation_prompts is not None and global_step % args.validation_steps == 0:
+ # if not args.train_text_encoder:
+ # # create pipeline
+ # text_encoder_one, text_encoder_two, text_encoder_three = load_text_encoders(
+ # text_encoder_cls_one, text_encoder_cls_two, text_encoder_cls_three
+ # )
+ # text_encoder_one.to(weight_dtype)
+ # text_encoder_two.to(weight_dtype)
+ pipeline = StableDiffusion3Pipeline.from_pretrained(
+ args.pretrained_model_name_or_path,
+ vae=vae,
+ text_encoder=accelerator.unwrap_model(text_encoder_one),
+ text_encoder_2=accelerator.unwrap_model(text_encoder_two),
+ text_encoder_3=accelerator.unwrap_model(text_encoder_three),
+ transformer=accelerator.unwrap_model(transformer),
+ revision=args.revision,
+ variant=args.variant,
+ torch_dtype=weight_dtype,
+ )#这里的pipeline是根据之前的训练数据更新过的，其中transformer是已经包含了lora层的
+ images = []
+ if args.validation_prompts and args.num_validation_images > 0:
+ pipeline_args = [{"prompt": prompt} for prompt in args.validation_prompts] # 假设是一个 prompt 列表
+ images = log_validation(
+ pipeline=pipeline,
+ args=args,
+ accelerator=accelerator,
+ pipeline_args=pipeline_args,
+ global_step=global_step,
+ torch_dtype=weight_dtype,
+ )
+ for i,image in enumerate(images):
+
+ # 保存图片，命名为 image_0.png, image_1.png ...
+ image_path = os.path.join(save_path, f"validation_image_{i}.png")
+ image.save(image_path)
+ images = None
+ del pipeline
+
+ logs = {
+ "loss": loss.detach().item(),
+ "raw_model_loss": raw_model_loss.detach().item(),
+ "ref_loss": raw_ref_loss.detach().item(),
+ "lr": lr_scheduler.get_last_lr()[0],
+ }
+ progress_bar.set_postfix(**logs)
+ accelerator.log(logs, step=global_step)
+
+ if global_step >= args.max_train_steps:
+ break
+ #-----------------------------------------------------------------
+ #每个周期里的所有batch都已经遍历完全，接下来的代码又回到了周期层面
+
+
+
+ # if not args.train_text_encoder:
+ # del text_encoder_one, text_encoder_two, text_encoder_three
+ # free_memory()#我个人还是感觉，训不训练文本编码部分，都不应该对text_encoder进行删除，释放显存，因为一直会需要使用
+ #--------------------------------------------------------------
+ #所有的周期都结束了，所有的训练也都结束了，接下来是保存以及进行最后一次的验证
+ # Save the lora layers
+ accelerator.wait_for_everyone()
+ if accelerator.is_main_process:
+ transformer = unwrap_model(transformer)
+ if args.upcast_before_saving:
+ transformer.to(torch.float32)
+ else:
+ transformer = transformer.to(weight_dtype)
+ transformer_lora_layers = get_peft_model_state_dict(transformer)
+
+ if args.train_text_encoder:
+ text_encoder_one = unwrap_model(text_encoder_one)
+ text_encoder_lora_layers = get_peft_model_state_dict(text_encoder_one.to(torch.float32))
+ text_encoder_two = unwrap_model(text_encoder_two)
+ text_encoder_2_lora_layers = get_peft_model_state_dict(text_encoder_two.to(torch.float32))
+ else:
+ text_encoder_lora_layers = None
+ text_encoder_2_lora_layers = None
+
+ StableDiffusion3Pipeline.save_lora_weights(
+ save_directory=args.output_dir,
+ transformer_lora_layers=transformer_lora_layers,
+ text_encoder_lora_layers=text_encoder_lora_layers,
+ text_encoder_2_lora_layers=text_encoder_2_lora_layers,
+ )#在训练结束之后会将lora的参数再次进行保存
+ # 💡 在加载最终验证 pipeline 之前，释放不再需要的训练模型以避免 OOM
+ del transformer, text_encoder_one, text_encoder_two, text_encoder_three, vae
+ torch.cuda.empty_cache()
+
+
+ # # Final inference
+ # # Load previous pipeline
+ # pipeline = StableDiffusion3Pipeline.from_pretrained(
+ # args.pretrained_model_name_or_path,
+ # revision=args.revision,
+ # variant=args.variant,
+ # torch_dtype=weight_dtype,
+ # )
+ # # load attention processors
+ # pipeline.load_lora_weights(args.output_dir)
+
+ # # run inference
+ # images = []
+ # if args.validation_prompts and args.num_validation_images > 0:
+ # pipeline_args = [{"prompt": prompt} for prompt in args.validation_prompts] # 假设是一个 prompt 列表
+ # images = log_validation(
+ # pipeline=pipeline,
+ # args=args,
+ # accelerator=accelerator,
+ # pipeline_args=pipeline_args,
+ # epoch=epoch,
+ # torch_dtype=weight_dtype,
+ # )
+ # for i,image in enumerate(images):
+
+ # # 保存图片，命名为 image_0.png, image_1.png ...
+ # image_path = os.path.join(args.output_dir, f"final_validation_image_{i}.png")
+ # image.save(image_path)
+
+ if args.push_to_hub:
+ save_model_card(
+ repo_id,
+ images=images,
+ base_model=args.pretrained_model_name_or_path,
+ instance_prompt=args.instance_prompt,
+ validation_prompts=args.validation_prompts,
+ train_text_encoder=args.train_text_encoder,
+ repo_folder=args.output_dir,
+ )
+ upload_folder(
+ repo_id=repo_id,
+ folder_path=args.output_dir,
+ commit_message="End of training",
+ ignore_patterns=["step_*", "epoch_*"],
+ )
+
+ accelerator.end_training()
+
+
+if __name__ == "__main__":
+ args = parse_args()
+ main(args)