update checkpoint

model-00001-of-00008.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:15c2451e4e0dd5caae61e91de67ea0dac7b554c4e2b39d54e67ffbe232460063
-size 4974581824
+oid sha256:6ef2c97b7b439a6e09dfe0587a490944d9f2c0e5bb0d13c1f26596465ff926d9
+size 4974580360

model-00002-of-00008.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:351e987e5f1f28124a8c839085aeb3111e8f956393d3092439e07c7a285a5d90
-size 4982995648
+oid sha256:d62a815ebcb470126c3c9921dffc96ffb893e0aaa5b79dbf773ce72d79b227a8
+size 4982994176

model-00003-of-00008.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ff93f061698e4a8bbc042d9f243efd9647b19889f0e5f87188cd6bfc0400f839
-size 4982995728
+oid sha256:73314e1b3507a3d730b3e9e1a4adb28df0e24090a3746934cf048db2faa50c9b
+size 4982994248

model-00004-of-00008.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:54c8f6eca491986fb61914029014d80a7813c1538807a55296530c6337fe2829
-size 4982995728
+oid sha256:b1bc455f58220ebe42f6d1008bfb2ef7a059837b6dadcc9bb6db033dc3235f00
+size 4982994248

model-00005-of-00008.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4218cb8e5bda353d4491a70054b2e8f384f00c55bcc453807309982760fc48fd
-size 4982995728
+oid sha256:aff0a61b3ac83772b56ebd28c8822abe0f86d8e2fdbd730d68a464ed1be3a1bd
+size 4982994248

model-00006-of-00008.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:23f2f3123c49dc4fe16fa44f6ee3dd388b18490fa05966f9dd21e806e3cbd22d
-size 4950060832
+oid sha256:4c2081cf569e6d8627032f9a41681aef8630583467a0215e02c3faccb426654b
+size 4950060344

model-00007-of-00008.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8d6eec2f2b61a1eb39dfb7703231639e16394bba2348108f1ed970268abac86e
+oid sha256:5c55becad2b16402294a2c1e67d00ef205f1f4f255aadadeaf20bc17b8a6a56f
 size 4945866712

model-00008-of-00008.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6df8c2486c4d21b20c5b2a72a699b47fd8ac62199ba9ae69db32054ef4fab1a2
+oid sha256:34e589fb9b2ed19cf98735c3b51a709dceb14e7d1b1503fc9601a150453ef265
 size 1783098344

model.safetensors.index.json

@@ -1,6 +1,6 @@
 {
   "metadata": {
-    "total_size": 36585351552
+    "total_size": 36585347456
   },
   "weight_map": {
     "lm_head.weight": "model-00008-of-00008.safetensors",
@@ -523,7 +523,6 @@
     "model.layers.0.cross_attn.dense.weight": "model-00001-of-00008.safetensors",
     "model.layers.0.cross_attn.key_value.weight": "model-00001-of-00008.safetensors",
     "model.layers.0.cross_attn.query.weight": "model-00001-of-00008.safetensors",
-    "model.layers.0.cross_attn.rotary_emb.inv_freq": "model-00001-of-00008.safetensors",
     "model.layers.0.input_layernorm.weight": "model-00001-of-00008.safetensors",
     "model.layers.0.mlp.language_mlp.down_proj.weight": "model-00001-of-00008.safetensors",
     "model.layers.0.mlp.language_mlp.gate_proj.weight": "model-00001-of-00008.safetensors",
@@ -541,7 +540,6 @@
     "model.layers.1.cross_attn.dense.weight": "model-00001-of-00008.safetensors",
     "model.layers.1.cross_attn.key_value.weight": "model-00001-of-00008.safetensors",
     "model.layers.1.cross_attn.query.weight": "model-00001-of-00008.safetensors",
-    "model.layers.1.cross_attn.rotary_emb.inv_freq": "model-00001-of-00008.safetensors",
     "model.layers.1.input_layernorm.weight": "model-00001-of-00008.safetensors",
     "model.layers.1.mlp.language_mlp.down_proj.weight": "model-00001-of-00008.safetensors",
     "model.layers.1.mlp.language_mlp.gate_proj.weight": "model-00001-of-00008.safetensors",
@@ -559,7 +557,6 @@
     "model.layers.10.cross_attn.dense.weight": "model-00002-of-00008.safetensors",
     "model.layers.10.cross_attn.key_value.weight": "model-00002-of-00008.safetensors",
     "model.layers.10.cross_attn.query.weight": "model-00002-of-00008.safetensors",
-    "model.layers.10.cross_attn.rotary_emb.inv_freq": "model-00002-of-00008.safetensors",
     "model.layers.10.input_layernorm.weight": "model-00002-of-00008.safetensors",
     "model.layers.10.mlp.language_mlp.down_proj.weight": "model-00002-of-00008.safetensors",
     "model.layers.10.mlp.language_mlp.gate_proj.weight": "model-00002-of-00008.safetensors",
@@ -577,7 +574,6 @@
     "model.layers.11.cross_attn.dense.weight": "model-00002-of-00008.safetensors",
     "model.layers.11.cross_attn.key_value.weight": "model-00002-of-00008.safetensors",
     "model.layers.11.cross_attn.query.weight": "model-00002-of-00008.safetensors",
-    "model.layers.11.cross_attn.rotary_emb.inv_freq": "model-00002-of-00008.safetensors",
     "model.layers.11.input_layernorm.weight": "model-00003-of-00008.safetensors",
     "model.layers.11.mlp.language_mlp.down_proj.weight": "model-00002-of-00008.safetensors",
     "model.layers.11.mlp.language_mlp.gate_proj.weight": "model-00002-of-00008.safetensors",
@@ -595,7 +591,6 @@
     "model.layers.12.cross_attn.dense.weight": "model-00003-of-00008.safetensors",
     "model.layers.12.cross_attn.key_value.weight": "model-00003-of-00008.safetensors",
     "model.layers.12.cross_attn.query.weight": "model-00003-of-00008.safetensors",
-    "model.layers.12.cross_attn.rotary_emb.inv_freq": "model-00003-of-00008.safetensors",
     "model.layers.12.input_layernorm.weight": "model-00003-of-00008.safetensors",
     "model.layers.12.mlp.language_mlp.down_proj.weight": "model-00003-of-00008.safetensors",
     "model.layers.12.mlp.language_mlp.gate_proj.weight": "model-00003-of-00008.safetensors",
@@ -613,7 +608,6 @@
     "model.layers.13.cross_attn.dense.weight": "model-00003-of-00008.safetensors",
     "model.layers.13.cross_attn.key_value.weight": "model-00003-of-00008.safetensors",
     "model.layers.13.cross_attn.query.weight": "model-00003-of-00008.safetensors",
-    "model.layers.13.cross_attn.rotary_emb.inv_freq": "model-00003-of-00008.safetensors",
     "model.layers.13.input_layernorm.weight": "model-00003-of-00008.safetensors",
     "model.layers.13.mlp.language_mlp.down_proj.weight": "model-00003-of-00008.safetensors",
     "model.layers.13.mlp.language_mlp.gate_proj.weight": "model-00003-of-00008.safetensors",
@@ -631,7 +625,6 @@
     "model.layers.14.cross_attn.dense.weight": "model-00003-of-00008.safetensors",
     "model.layers.14.cross_attn.key_value.weight": "model-00003-of-00008.safetensors",
     "model.layers.14.cross_attn.query.weight": "model-00003-of-00008.safetensors",
-    "model.layers.14.cross_attn.rotary_emb.inv_freq": "model-00003-of-00008.safetensors",
     "model.layers.14.input_layernorm.weight": "model-00003-of-00008.safetensors",
     "model.layers.14.mlp.language_mlp.down_proj.weight": "model-00003-of-00008.safetensors",
     "model.layers.14.mlp.language_mlp.gate_proj.weight": "model-00003-of-00008.safetensors",
@@ -649,7 +642,6 @@
     "model.layers.15.cross_attn.dense.weight": "model-00003-of-00008.safetensors",
     "model.layers.15.cross_attn.key_value.weight": "model-00003-of-00008.safetensors",
     "model.layers.15.cross_attn.query.weight": "model-00003-of-00008.safetensors",
-    "model.layers.15.cross_attn.rotary_emb.inv_freq": "model-00003-of-00008.safetensors",
     "model.layers.15.input_layernorm.weight": "model-00003-of-00008.safetensors",
     "model.layers.15.mlp.language_mlp.down_proj.weight": "model-00003-of-00008.safetensors",
     "model.layers.15.mlp.language_mlp.gate_proj.weight": "model-00003-of-00008.safetensors",
@@ -667,7 +659,6 @@
     "model.layers.16.cross_attn.dense.weight": "model-00003-of-00008.safetensors",
     "model.layers.16.cross_attn.key_value.weight": "model-00003-of-00008.safetensors",
     "model.layers.16.cross_attn.query.weight": "model-00003-of-00008.safetensors",
-    "model.layers.16.cross_attn.rotary_emb.inv_freq": "model-00003-of-00008.safetensors",
     "model.layers.16.input_layernorm.weight": "model-00003-of-00008.safetensors",
     "model.layers.16.mlp.language_mlp.down_proj.weight": "model-00003-of-00008.safetensors",
     "model.layers.16.mlp.language_mlp.gate_proj.weight": "model-00003-of-00008.safetensors",
@@ -685,7 +676,6 @@
     "model.layers.17.cross_attn.dense.weight": "model-00003-of-00008.safetensors",
     "model.layers.17.cross_attn.key_value.weight": "model-00003-of-00008.safetensors",
     "model.layers.17.cross_attn.query.weight": "model-00003-of-00008.safetensors",
-    "model.layers.17.cross_attn.rotary_emb.inv_freq": "model-00003-of-00008.safetensors",
     "model.layers.17.input_layernorm.weight": "model-00004-of-00008.safetensors",
     "model.layers.17.mlp.language_mlp.down_proj.weight": "model-00003-of-00008.safetensors",
     "model.layers.17.mlp.language_mlp.gate_proj.weight": "model-00003-of-00008.safetensors",
@@ -703,7 +693,6 @@
     "model.layers.18.cross_attn.dense.weight": "model-00004-of-00008.safetensors",
     "model.layers.18.cross_attn.key_value.weight": "model-00004-of-00008.safetensors",
     "model.layers.18.cross_attn.query.weight": "model-00004-of-00008.safetensors",
-    "model.layers.18.cross_attn.rotary_emb.inv_freq": "model-00004-of-00008.safetensors",
     "model.layers.18.input_layernorm.weight": "model-00004-of-00008.safetensors",
     "model.layers.18.mlp.language_mlp.down_proj.weight": "model-00004-of-00008.safetensors",
     "model.layers.18.mlp.language_mlp.gate_proj.weight": "model-00004-of-00008.safetensors",
@@ -721,7 +710,6 @@
     "model.layers.19.cross_attn.dense.weight": "model-00004-of-00008.safetensors",
     "model.layers.19.cross_attn.key_value.weight": "model-00004-of-00008.safetensors",
     "model.layers.19.cross_attn.query.weight": "model-00004-of-00008.safetensors",
-    "model.layers.19.cross_attn.rotary_emb.inv_freq": "model-00004-of-00008.safetensors",
     "model.layers.19.input_layernorm.weight": "model-00004-of-00008.safetensors",
     "model.layers.19.mlp.language_mlp.down_proj.weight": "model-00004-of-00008.safetensors",
     "model.layers.19.mlp.language_mlp.gate_proj.weight": "model-00004-of-00008.safetensors",
@@ -739,7 +727,6 @@
     "model.layers.2.cross_attn.dense.weight": "model-00001-of-00008.safetensors",
     "model.layers.2.cross_attn.key_value.weight": "model-00001-of-00008.safetensors",
     "model.layers.2.cross_attn.query.weight": "model-00001-of-00008.safetensors",
-    "model.layers.2.cross_attn.rotary_emb.inv_freq": "model-00001-of-00008.safetensors",
     "model.layers.2.input_layernorm.weight": "model-00001-of-00008.safetensors",
     "model.layers.2.mlp.language_mlp.down_proj.weight": "model-00001-of-00008.safetensors",
     "model.layers.2.mlp.language_mlp.gate_proj.weight": "model-00001-of-00008.safetensors",
@@ -757,7 +744,6 @@
     "model.layers.20.cross_attn.dense.weight": "model-00004-of-00008.safetensors",
     "model.layers.20.cross_attn.key_value.weight": "model-00004-of-00008.safetensors",
     "model.layers.20.cross_attn.query.weight": "model-00004-of-00008.safetensors",
-    "model.layers.20.cross_attn.rotary_emb.inv_freq": "model-00004-of-00008.safetensors",
     "model.layers.20.input_layernorm.weight": "model-00004-of-00008.safetensors",
     "model.layers.20.mlp.language_mlp.down_proj.weight": "model-00004-of-00008.safetensors",
     "model.layers.20.mlp.language_mlp.gate_proj.weight": "model-00004-of-00008.safetensors",
@@ -775,7 +761,6 @@
     "model.layers.21.cross_attn.dense.weight": "model-00004-of-00008.safetensors",
     "model.layers.21.cross_attn.key_value.weight": "model-00004-of-00008.safetensors",
     "model.layers.21.cross_attn.query.weight": "model-00004-of-00008.safetensors",
-    "model.layers.21.cross_attn.rotary_emb.inv_freq": "model-00004-of-00008.safetensors",
     "model.layers.21.input_layernorm.weight": "model-00004-of-00008.safetensors",
     "model.layers.21.mlp.language_mlp.down_proj.weight": "model-00004-of-00008.safetensors",
     "model.layers.21.mlp.language_mlp.gate_proj.weight": "model-00004-of-00008.safetensors",
@@ -793,7 +778,6 @@
     "model.layers.22.cross_attn.dense.weight": "model-00004-of-00008.safetensors",
     "model.layers.22.cross_attn.key_value.weight": "model-00004-of-00008.safetensors",
     "model.layers.22.cross_attn.query.weight": "model-00004-of-00008.safetensors",
-    "model.layers.22.cross_attn.rotary_emb.inv_freq": "model-00004-of-00008.safetensors",
     "model.layers.22.input_layernorm.weight": "model-00004-of-00008.safetensors",
     "model.layers.22.mlp.language_mlp.down_proj.weight": "model-00004-of-00008.safetensors",
     "model.layers.22.mlp.language_mlp.gate_proj.weight": "model-00004-of-00008.safetensors",
@@ -811,7 +795,6 @@
     "model.layers.23.cross_attn.dense.weight": "model-00004-of-00008.safetensors",
     "model.layers.23.cross_attn.key_value.weight": "model-00004-of-00008.safetensors",
     "model.layers.23.cross_attn.query.weight": "model-00004-of-00008.safetensors",
-    "model.layers.23.cross_attn.rotary_emb.inv_freq": "model-00004-of-00008.safetensors",
     "model.layers.23.input_layernorm.weight": "model-00005-of-00008.safetensors",
     "model.layers.23.mlp.language_mlp.down_proj.weight": "model-00004-of-00008.safetensors",
     "model.layers.23.mlp.language_mlp.gate_proj.weight": "model-00004-of-00008.safetensors",
@@ -829,7 +812,6 @@
     "model.layers.24.cross_attn.dense.weight": "model-00005-of-00008.safetensors",
     "model.layers.24.cross_attn.key_value.weight": "model-00005-of-00008.safetensors",
     "model.layers.24.cross_attn.query.weight": "model-00005-of-00008.safetensors",
-    "model.layers.24.cross_attn.rotary_emb.inv_freq": "model-00005-of-00008.safetensors",
     "model.layers.24.input_layernorm.weight": "model-00005-of-00008.safetensors",
     "model.layers.24.mlp.language_mlp.down_proj.weight": "model-00005-of-00008.safetensors",
     "model.layers.24.mlp.language_mlp.gate_proj.weight": "model-00005-of-00008.safetensors",
@@ -847,7 +829,6 @@
     "model.layers.25.cross_attn.dense.weight": "model-00005-of-00008.safetensors",
     "model.layers.25.cross_attn.key_value.weight": "model-00005-of-00008.safetensors",
     "model.layers.25.cross_attn.query.weight": "model-00005-of-00008.safetensors",
-    "model.layers.25.cross_attn.rotary_emb.inv_freq": "model-00005-of-00008.safetensors",
     "model.layers.25.input_layernorm.weight": "model-00005-of-00008.safetensors",
     "model.layers.25.mlp.language_mlp.down_proj.weight": "model-00005-of-00008.safetensors",
     "model.layers.25.mlp.language_mlp.gate_proj.weight": "model-00005-of-00008.safetensors",
@@ -865,7 +846,6 @@
     "model.layers.26.cross_attn.dense.weight": "model-00005-of-00008.safetensors",
     "model.layers.26.cross_attn.key_value.weight": "model-00005-of-00008.safetensors",
     "model.layers.26.cross_attn.query.weight": "model-00005-of-00008.safetensors",
-    "model.layers.26.cross_attn.rotary_emb.inv_freq": "model-00005-of-00008.safetensors",
     "model.layers.26.input_layernorm.weight": "model-00005-of-00008.safetensors",
     "model.layers.26.mlp.language_mlp.down_proj.weight": "model-00005-of-00008.safetensors",
     "model.layers.26.mlp.language_mlp.gate_proj.weight": "model-00005-of-00008.safetensors",
@@ -883,7 +863,6 @@
     "model.layers.27.cross_attn.dense.weight": "model-00005-of-00008.safetensors",
     "model.layers.27.cross_attn.key_value.weight": "model-00005-of-00008.safetensors",
     "model.layers.27.cross_attn.query.weight": "model-00005-of-00008.safetensors",
-    "model.layers.27.cross_attn.rotary_emb.inv_freq": "model-00005-of-00008.safetensors",
     "model.layers.27.input_layernorm.weight": "model-00005-of-00008.safetensors",
     "model.layers.27.mlp.language_mlp.down_proj.weight": "model-00005-of-00008.safetensors",
     "model.layers.27.mlp.language_mlp.gate_proj.weight": "model-00005-of-00008.safetensors",
@@ -901,7 +880,6 @@
     "model.layers.28.cross_attn.dense.weight": "model-00005-of-00008.safetensors",
     "model.layers.28.cross_attn.key_value.weight": "model-00005-of-00008.safetensors",
     "model.layers.28.cross_attn.query.weight": "model-00005-of-00008.safetensors",
-    "model.layers.28.cross_attn.rotary_emb.inv_freq": "model-00005-of-00008.safetensors",
     "model.layers.28.input_layernorm.weight": "model-00005-of-00008.safetensors",
     "model.layers.28.mlp.language_mlp.down_proj.weight": "model-00005-of-00008.safetensors",
     "model.layers.28.mlp.language_mlp.gate_proj.weight": "model-00005-of-00008.safetensors",
@@ -919,7 +897,6 @@
     "model.layers.29.cross_attn.dense.weight": "model-00005-of-00008.safetensors",
     "model.layers.29.cross_attn.key_value.weight": "model-00005-of-00008.safetensors",
     "model.layers.29.cross_attn.query.weight": "model-00005-of-00008.safetensors",
-    "model.layers.29.cross_attn.rotary_emb.inv_freq": "model-00005-of-00008.safetensors",
     "model.layers.29.input_layernorm.weight": "model-00006-of-00008.safetensors",
     "model.layers.29.mlp.language_mlp.down_proj.weight": "model-00005-of-00008.safetensors",
     "model.layers.29.mlp.language_mlp.gate_proj.weight": "model-00005-of-00008.safetensors",
@@ -937,7 +914,6 @@
     "model.layers.3.cross_attn.dense.weight": "model-00001-of-00008.safetensors",
     "model.layers.3.cross_attn.key_value.weight": "model-00001-of-00008.safetensors",
     "model.layers.3.cross_attn.query.weight": "model-00001-of-00008.safetensors",
-    "model.layers.3.cross_attn.rotary_emb.inv_freq": "model-00001-of-00008.safetensors",
     "model.layers.3.input_layernorm.weight": "model-00001-of-00008.safetensors",
     "model.layers.3.mlp.language_mlp.down_proj.weight": "model-00001-of-00008.safetensors",
     "model.layers.3.mlp.language_mlp.gate_proj.weight": "model-00001-of-00008.safetensors",
@@ -955,7 +931,6 @@
     "model.layers.30.cross_attn.dense.weight": "model-00006-of-00008.safetensors",
     "model.layers.30.cross_attn.key_value.weight": "model-00006-of-00008.safetensors",
     "model.layers.30.cross_attn.query.weight": "model-00006-of-00008.safetensors",
-    "model.layers.30.cross_attn.rotary_emb.inv_freq": "model-00006-of-00008.safetensors",
     "model.layers.30.input_layernorm.weight": "model-00006-of-00008.safetensors",
     "model.layers.30.mlp.language_mlp.down_proj.weight": "model-00006-of-00008.safetensors",
     "model.layers.30.mlp.language_mlp.gate_proj.weight": "model-00006-of-00008.safetensors",
@@ -973,7 +948,6 @@
     "model.layers.31.cross_attn.dense.weight": "model-00006-of-00008.safetensors",
     "model.layers.31.cross_attn.key_value.weight": "model-00006-of-00008.safetensors",
     "model.layers.31.cross_attn.query.weight": "model-00006-of-00008.safetensors",
-    "model.layers.31.cross_attn.rotary_emb.inv_freq": "model-00006-of-00008.safetensors",
     "model.layers.31.input_layernorm.weight": "model-00006-of-00008.safetensors",
     "model.layers.31.mlp.language_mlp.down_proj.weight": "model-00006-of-00008.safetensors",
     "model.layers.31.mlp.language_mlp.gate_proj.weight": "model-00006-of-00008.safetensors",
@@ -991,7 +965,6 @@
     "model.layers.4.cross_attn.dense.weight": "model-00001-of-00008.safetensors",
     "model.layers.4.cross_attn.key_value.weight": "model-00001-of-00008.safetensors",
     "model.layers.4.cross_attn.query.weight": "model-00001-of-00008.safetensors",
-    "model.layers.4.cross_attn.rotary_emb.inv_freq": "model-00001-of-00008.safetensors",
     "model.layers.4.input_layernorm.weight": "model-00001-of-00008.safetensors",
     "model.layers.4.mlp.language_mlp.down_proj.weight": "model-00001-of-00008.safetensors",
     "model.layers.4.mlp.language_mlp.gate_proj.weight": "model-00001-of-00008.safetensors",
@@ -1009,7 +982,6 @@
     "model.layers.5.cross_attn.dense.weight": "model-00001-of-00008.safetensors",
     "model.layers.5.cross_attn.key_value.weight": "model-00001-of-00008.safetensors",
     "model.layers.5.cross_attn.query.weight": "model-00001-of-00008.safetensors",
-    "model.layers.5.cross_attn.rotary_emb.inv_freq": "model-00001-of-00008.safetensors",
     "model.layers.5.input_layernorm.weight": "model-00002-of-00008.safetensors",
     "model.layers.5.mlp.language_mlp.down_proj.weight": "model-00001-of-00008.safetensors",
     "model.layers.5.mlp.language_mlp.gate_proj.weight": "model-00001-of-00008.safetensors",
@@ -1027,7 +999,6 @@
     "model.layers.6.cross_attn.dense.weight": "model-00002-of-00008.safetensors",
     "model.layers.6.cross_attn.key_value.weight": "model-00002-of-00008.safetensors",
     "model.layers.6.cross_attn.query.weight": "model-00002-of-00008.safetensors",
-    "model.layers.6.cross_attn.rotary_emb.inv_freq": "model-00002-of-00008.safetensors",
     "model.layers.6.input_layernorm.weight": "model-00002-of-00008.safetensors",
     "model.layers.6.mlp.language_mlp.down_proj.weight": "model-00002-of-00008.safetensors",
     "model.layers.6.mlp.language_mlp.gate_proj.weight": "model-00002-of-00008.safetensors",
@@ -1045,7 +1016,6 @@
     "model.layers.7.cross_attn.dense.weight": "model-00002-of-00008.safetensors",
     "model.layers.7.cross_attn.key_value.weight": "model-00002-of-00008.safetensors",
     "model.layers.7.cross_attn.query.weight": "model-00002-of-00008.safetensors",
-    "model.layers.7.cross_attn.rotary_emb.inv_freq": "model-00002-of-00008.safetensors",
     "model.layers.7.input_layernorm.weight": "model-00002-of-00008.safetensors",
     "model.layers.7.mlp.language_mlp.down_proj.weight": "model-00002-of-00008.safetensors",
     "model.layers.7.mlp.language_mlp.gate_proj.weight": "model-00002-of-00008.safetensors",
@@ -1063,7 +1033,6 @@
     "model.layers.8.cross_attn.dense.weight": "model-00002-of-00008.safetensors",
     "model.layers.8.cross_attn.key_value.weight": "model-00002-of-00008.safetensors",
     "model.layers.8.cross_attn.query.weight": "model-00002-of-00008.safetensors",
-    "model.layers.8.cross_attn.rotary_emb.inv_freq": "model-00002-of-00008.safetensors",
     "model.layers.8.input_layernorm.weight": "model-00002-of-00008.safetensors",
     "model.layers.8.mlp.language_mlp.down_proj.weight": "model-00002-of-00008.safetensors",
     "model.layers.8.mlp.language_mlp.gate_proj.weight": "model-00002-of-00008.safetensors",
@@ -1081,7 +1050,6 @@
     "model.layers.9.cross_attn.dense.weight": "model-00002-of-00008.safetensors",
     "model.layers.9.cross_attn.key_value.weight": "model-00002-of-00008.safetensors",
     "model.layers.9.cross_attn.query.weight": "model-00002-of-00008.safetensors",
-    "model.layers.9.cross_attn.rotary_emb.inv_freq": "model-00002-of-00008.safetensors",
     "model.layers.9.input_layernorm.weight": "model-00002-of-00008.safetensors",
     "model.layers.9.mlp.language_mlp.down_proj.weight": "model-00002-of-00008.safetensors",
     "model.layers.9.mlp.language_mlp.gate_proj.weight": "model-00002-of-00008.safetensors",

modeling_cogagent.py

@@ -1,965 +1,974 @@
-"""largely copy from llama and adapt for CogAgent"""
-import warnings
-from typing import TYPE_CHECKING, Optional, Tuple, List, Union, Literal, Dict, Any
-
-import math
-import torch
-from torch import nn
-from torch.nn import CrossEntropyLoss
-from torchvision import transforms
-from einops import rearrange
-
-from transformers import PreTrainedModel, PreTrainedTokenizer
-from transformers.utils.logging import get_logger
-from transformers.activations import ACT2FN
-from transformers.modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast
-
-from .configuration_cogagent import CogAgentConfig
-# from .util import FastRotaryEmbedding
-from torch.nn import functional as F
-from .visual import EVA2CLIPModel
-from .cross_visual import CrossVisionModel
-
-if TYPE_CHECKING:
-    from transformers.utils import ModelOutput
-
-logger = get_logger(__name__)
-
-LANGUAGE_TOKEN_TYPE = 0
-VISION_TOKEN_TYPE = 1
-
-
-# Copied from transformers.models.bart.modeling_bart._make_causal_mask
-def _make_causal_mask(
-        input_ids_shape: torch.Size, dtype: torch.dtype, device: torch.device, past_key_values_length: int = 0
-):
-    """
-    Make causal mask used for bi-directional self-attention.
-    """
-    bsz, tgt_len = input_ids_shape
-    mask = torch.full((tgt_len, tgt_len), torch.finfo(dtype).min, device=device)
-    mask_cond = torch.arange(mask.size(-1), device=device)
-    mask.masked_fill_(mask_cond < (mask_cond + 1).view(mask.size(-1), 1), 0)
-    mask = mask.to(dtype)
-
-    if past_key_values_length > 0:
-        mask = torch.cat([torch.zeros(tgt_len, past_key_values_length, dtype=dtype, device=device), mask], dim=-1)
-    return mask[None, None, :, :].expand(bsz, 1, tgt_len, tgt_len + past_key_values_length)
-
-
-# Copied from transformers.models.bart.modeling_bart._expand_mask
-def _expand_mask(mask: torch.Tensor, dtype: torch.dtype, tgt_len: Optional[int] = None):
-    """
-    Expands attention_mask from `[bsz, seq_len]` to `[bsz, 1, tgt_seq_len, src_seq_len]`.
-    """
-    bsz, src_len = mask.size()
-    tgt_len = tgt_len if tgt_len is not None else src_len
-
-    expanded_mask = mask[:, None, None, :].expand(bsz, 1, tgt_len, src_len).to(dtype)
-
-    inverted_mask = 1.0 - expanded_mask
-
-    return inverted_mask.masked_fill(inverted_mask.to(torch.bool), torch.finfo(dtype).min)
-
-
-class RMSNorm(nn.Module):
-    def __init__(self, hidden_size, eps=1e-6):
-        super().__init__()
-        self.weight = nn.Parameter(torch.ones(hidden_size))
-        self.variance_epsilon = eps
-
-    def forward(self, hidden_states):
-        input_dtype = hidden_states.dtype
-        hidden_states = hidden_states.to(torch.float32)
-        variance = hidden_states.pow(2).mean(-1, keepdim=True)
-        hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
-        return (self.weight * hidden_states).to(input_dtype)
-
-
-class MLP(nn.Module):
-    def __init__(self, config):
-        super().__init__()
-        self.hidden_size = config.hidden_size
-        self.intermediate_size = config.intermediate_size
-        self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
-        self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
-        self.down_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=False)
-        self.act_fn = ACT2FN[config.hidden_act]
-
-    def forward(self, x):
-        down_proj = self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))
-        return down_proj
-
-
-def get_expert_mask(token_type_ids: "torch.LongTensor(B, L)") -> "[torch.BoolTensor(B, L), torch.BoolTensor(B, L)]":
-    vision_token_mask = torch.zeros_like(token_type_ids, dtype=torch.bool)
-    vision_token_mask[:, :-1] = (token_type_ids[:, :-1] == VISION_TOKEN_TYPE) & (token_type_ids[:, 1:] == VISION_TOKEN_TYPE)
-    language_token_mask = ~vision_token_mask
-    return vision_token_mask, language_token_mask
-
-
-class VisionExpertMLP(nn.Module):
-    def __init__(self, config):
-        super().__init__()
-        self.language_mlp = MLP(config)
-        self.vision_mlp = MLP(config)
-
-    def forward(self, hidden_states: "torch.Tensor(B, L, D)", token_type_ids: "torch.LongTensor(B, L)"):
-        output = torch.empty(hidden_states.shape, dtype=hidden_states.dtype, device=hidden_states.device)
-        vision_token_mask, language_token_mask = get_expert_mask(token_type_ids)
-        output[vision_token_mask] = self.vision_mlp(hidden_states[vision_token_mask])
-        output[language_token_mask] = self.language_mlp(hidden_states[language_token_mask])
-        return output
-
-
-def attention_fn(
-        query_layer: "torch.tensor(B, H, L, HD)",
-        key_layer: "torch.tensor(B, H, L, HD)",
-        value_layer: "torch.tensor(B, H, L, HD)",
-        attention_mask: "torch.tensor(B, H, L, HD)",
-        *,
-        scaling_attention_score: bool = True,
-        attention_dropout: nn.Module = None
-):
-    attention_mask_bool = (attention_mask == 0)
-    is_low_triangle = (attention_mask_bool == torch.ones_like(attention_mask_bool, dtype=torch.float).tril()).all()
-    is_full = (attention_mask_bool > 0).all()
-    if not (int(torch.__version__.split('.')[0]) >= 2):
-        warnings.warn("It's recommended to use torch2.0 or higher.")
-    if int(torch.__version__.split('.')[0]) >= 2 and scaling_attention_score and (is_full or is_low_triangle):
-        dropout_p = 0. if attention_dropout is None or not attention_dropout.training else attention_dropout.p
-        return torch.nn.functional.scaled_dot_product_attention(
-            query_layer, key_layer, value_layer,
-            attn_mask=None,
-            dropout_p=dropout_p,
-            is_causal=not is_full
-        )
-    else:
-        if scaling_attention_score:
-            query_layer = query_layer / math.sqrt(query_layer.shape[-1])
-        attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2))
-        attention_scores = attention_scores + attention_mask
-        attention_scores = nn.functional.softmax(attention_scores, dim=-1, dtype=torch.float32).to(query_layer.dtype)
-        if attention_dropout is not None:
-            attention_scores = attention_dropout(attention_scores)
-        context_layer = torch.matmul(attention_scores, value_layer)
-        return context_layer
-
-class RotaryEmbedding(torch.nn.Module):
-    def __init__(self, dim, max_position_embeddings=2048, base=10000, device=None):
-        super().__init__()
-
-        self.dim = dim
-        self.max_position_embeddings = max_position_embeddings
-        self.base = base
-        inv_freq = self._compute_inv_freq(device)
-        self.register_buffer("inv_freq", inv_freq)
-        self.max_seq_len_cached = 0
-
-    def _compute_inv_freq(self, device=None):
-        return 1.0 / (
-                self.base
-                ** (torch.arange(0, self.dim, 2, device=device) / self.dim)
-        )
-
-    def _set_cos_sin_cache(self, seq_len, device, dtype):
-        self.max_seq_len_cached = seq_len
-        t = torch.arange(self.max_seq_len_cached, device=device, dtype=self.inv_freq.dtype)
-
-        freqs = torch.einsum("i,j->ij", t, self.inv_freq)
-        # Different from paper, but it uses a different permutation in order to obtain the same calculation
-        emb = torch.cat((freqs, freqs), dim=-1)
-        self.register_buffer("cos_cached", emb.cos()[:, None, :].to(dtype), persistent=False)
-        self.register_buffer("sin_cached", emb.sin()[:, None, :].to(dtype), persistent=False)
-
-    def forward(self, x, seq_len):
-        # x: [bs, num_attention_heads, seq_len, head_size]
-        if seq_len > self.max_seq_len_cached:
-            self._set_cos_sin_cache(seq_len=seq_len, device=x.device, dtype=x.dtype)
-
-        return (
-            self.cos_cached[:seq_len, ...].to(dtype=x.dtype),
-            self.sin_cached[:seq_len, ...].to(dtype=x.dtype),
-        )
-
-
-def rotate_half(x):
-    x1, x2 = x[..., :x.shape[-1] // 2], x[..., x.shape[-1] // 2:]
-    return torch.cat((-x2, x1), dim=x1.ndim - 1)
-
-
-def apply_rotary_pos_emb_index_bhs(q, k, cos, sin, position_id):
-    # batch_size, num_head, seq_len, hidden_size
-    cos, sin = F.embedding(position_id, cos.squeeze(1)).unsqueeze(1), \
-        F.embedding(position_id, sin.squeeze(1)).unsqueeze(1)
-    q, k = (q * cos) + (rotate_half(q) * sin), (k * cos) + (rotate_half(k) * sin)
-    return q, k
-
-class VisionExpertAttention(nn.Module):
-    def __init__(self, config):
-        super().__init__()
-        self.config = config
-        self.hidden_size = config.hidden_size
-        self.num_heads = config.num_attention_heads
-        self.head_dim = self.hidden_size // self.num_heads
-        self.max_position_embeddings = config.max_position_embeddings
-
-        self.rotary_emb = RotaryEmbedding(self.head_dim)
-        self.vision_expert_query_key_value = nn.Linear(self.hidden_size, self.hidden_size * 3, bias=False)
-        self.vision_expert_dense = nn.Linear(self.hidden_size, self.hidden_size, bias=False)
-        self.language_expert_query_key_value = nn.Linear(self.hidden_size, self.hidden_size * 3, bias=False)
-        self.language_expert_dense = nn.Linear(self.hidden_size, self.hidden_size, bias=False)
-
-    def _transpose_for_scores(self, tensor):
-        """Transpose a 3D tensor [B, L, H*HD] into a 4D tensor with size [B H L HD]."""
-        new_tensor_shape = tensor.size()[:-1] + (self.num_heads, self.head_dim)
-        tensor = tensor.view(*new_tensor_shape)
-        return tensor.permute(0, 2, 1, 3)
-
-    def forward(
-            self,
-            hidden_states: torch.Tensor,
-            token_type_ids: torch.LongTensor,
-            position_ids: torch.LongTensor,
-            attention_mask: Optional[torch.Tensor] = None,
-            past_key_value: Optional[Tuple[torch.Tensor]] = None,
-            output_attentions: bool = False,
-            use_cache: bool = False,
-    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
-        bsz, q_len, _ = hidden_states.size()
-        vision_token_mask, language_token_mask = get_expert_mask(token_type_ids)
-
-        shape = list(hidden_states.shape)
-        shape[-1] = shape[-1] * 3
-        mixed_raw_layer = torch.empty(shape, dtype=hidden_states.dtype, device=hidden_states.device)
-        mixed_raw_layer[vision_token_mask] = self.vision_expert_query_key_value(hidden_states[vision_token_mask])
-        mixed_raw_layer[language_token_mask] = self.language_expert_query_key_value(hidden_states[language_token_mask])
-
-        query_states, key_states, value_states = torch.split(mixed_raw_layer, self.hidden_size, dim=-1)
-        query_states = self._transpose_for_scores(query_states)  # B, H, L, HD
-        key_states = self._transpose_for_scores(key_states)  # B, H, L, HD
-        value_states = self._transpose_for_scores(value_states)  # B, H, L, HD
-
-        kv_seq_len = key_states.shape[-2]
-        if past_key_value is not None:
-            kv_seq_len += past_key_value[0].shape[-2]
-
-        cos, sin = self.rotary_emb(value_states, seq_len=position_ids.max() + 1)
-        query_states, key_states = apply_rotary_pos_emb_index_bhs(query_states, key_states, cos, sin, position_ids)
-
-        if past_key_value is not None:
-            key_states = torch.cat([past_key_value[0], key_states], dim=2)
-            value_states = torch.cat([past_key_value[1], value_states], dim=2)
-
-        past_key_value = (key_states, value_states) if use_cache else None
-
-        context_layer = attention_fn(
-            query_layer=query_states, key_layer=key_states, value_layer=value_states, attention_mask=attention_mask,
-            scaling_attention_score=True, attention_dropout=None)
-        if context_layer.size() != (bsz, self.num_heads, q_len, self.head_dim):
-            raise ValueError(
-                f"`attn_output` should be of size {(bsz, self.num_heads, q_len, self.head_dim)}, but is"
-                f" {context_layer.size()}"
-            )
-        context_layer = context_layer.transpose(1, 2).contiguous().reshape(bsz, q_len, self.hidden_size)
-
-        attn_output = torch.empty(context_layer.shape, dtype=hidden_states.dtype, device=hidden_states.device)
-        attn_output[vision_token_mask] = self.vision_expert_dense(context_layer[vision_token_mask])
-        attn_output[language_token_mask] = self.language_expert_dense(context_layer[language_token_mask])
-
-        if output_attentions:
-            warnings.warn("output_attentions is not implemented.")
-
-        return attn_output, None, past_key_value
-
-class CrossAttention(nn.Module):
-    def __init__(self, config):
-        super().__init__()
-        self.config = config
-        self.hidden_size = config.hidden_size
-        self.cross_hidden_size = config.cross_hidden_size
-        self.cross_compute_hidden_size = config.cross_compute_hidden_size
-        self.num_heads = config.num_attention_heads
-        self.head_dim = self.hidden_size // self.num_heads
-        self.cross_head_dim = self.cross_compute_hidden_size // self.num_heads
-        self.max_position_embeddings = config.max_position_embeddings
-
-        self.query = nn.Linear(self.hidden_size, self.cross_compute_hidden_size, bias=False)
-        self.key_value = nn.Linear(self.cross_hidden_size, self.cross_compute_hidden_size * 2, bias=False)
-        self.dense = nn.Linear(self.cross_compute_hidden_size, self.hidden_size, bias=False)
-
-    def _transpose_for_scores(self, tensor):
-        """Transpose a 3D tensor [B, L, H*HD] into a 4D tensor with size [B H L HD]."""
-        new_tensor_shape = tensor.size()[:-1] + (self.num_heads, self.cross_head_dim)
-        tensor = tensor.view(*new_tensor_shape)
-        return tensor.permute(0, 2, 1, 3)
-
-    def forward(
-            self,
-            hidden_states: torch.Tensor,
-            encoder_outputs: torch.LongTensor,
-            attention_mask: Optional[torch.Tensor] = None,
-            past_key_value: Optional[Tuple[torch.Tensor]] = None,
-            output_attentions: bool = False,
-            use_cache: bool = False,
-    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
-        bsz, q_len, _ = hidden_states.size()
-
-        shape = list(hidden_states.shape)
-        shape[-1] = shape[-1] * 3
-
-        mixed_query_layer = self.query(hidden_states)
-        if past_key_value is None:
-            mixed_x_layer = self.key_value(encoder_outputs)
-            mixed_key_layer, mixed_value_layer = torch.split(mixed_x_layer, self.cross_compute_hidden_size, dim=-1)
-            key_states = self._transpose_for_scores(mixed_key_layer)  # B, H, L, HD
-            value_states = self._transpose_for_scores(mixed_value_layer)  # B, H, L, HD
-        else:
-            key_states, value_states = past_key_value
-
-        query_states = self._transpose_for_scores(mixed_query_layer)  # B, H, L, HD
-
-        past_key_value = (key_states, value_states) if use_cache else None
-
-        context_layer = attention_fn(
-            query_layer=query_states, key_layer=key_states, value_layer=value_states, attention_mask=attention_mask,
-            scaling_attention_score=True, attention_dropout=None)
-        if context_layer.size() != (bsz, self.num_heads, q_len, self.cross_head_dim):
-            raise ValueError(
-                f"`cross_attn_output` should be of size {(bsz, self.num_heads, q_len, self.cross_head_dim)}, but is"
-                f" {context_layer.size()}"
-            )
-        context_layer = context_layer.transpose(1, 2).contiguous().reshape(bsz, q_len, self.cross_hidden_size)
-
-        attn_output = self.dense(context_layer)
-
-        if output_attentions:
-            warnings.warn("output_attentions is not implemented.")
-
-        return attn_output, None, past_key_value
-
-class CogAgentDecoderLayer(nn.Module):
-    def __init__(self, config):
-        super().__init__()
-        self.hidden_size = config.hidden_size
-        self.self_attn = VisionExpertAttention(config=config)
-        self.cross_attn = CrossAttention(config=config)
-        self.mlp = VisionExpertMLP(config)
-        self.input_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
-        self.post_attention_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
-        self.post_cross_attention_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
-
-    def forward(
-            self,
-            hidden_states: torch.Tensor,
-            encoder_outputs: torch.Tensor,
-            token_type_ids: torch.LongTensor,
-            position_ids: torch.LongTensor,
-            attention_mask: Optional[torch.Tensor] = None,
-            cross_attention_mask: Optional[torch.Tensor] = None,
-            past_key_value: Optional[Tuple[torch.Tensor]] = None,
-            output_attentions: Optional[bool] = False,
-            use_cache: Optional[bool] = False,
-    ) -> Tuple[torch.FloatTensor, Optional[Tuple[torch.FloatTensor, torch.FloatTensor]]]:
-        residual = hidden_states
-
-        hidden_states = self.input_layernorm(hidden_states)
-
-        # Self Attention
-        hidden_states, self_attn_weights, present_key_value = self.self_attn(
-            hidden_states=hidden_states,
-            token_type_ids=token_type_ids,
-            position_ids=position_ids,
-            attention_mask=attention_mask,
-            past_key_value=past_key_value[:2] if past_key_value is not None else None,
-            output_attentions=output_attentions,
-            use_cache=use_cache,
-        )
-        hidden_states = residual + hidden_states
-
-        cross_input = self.post_cross_attention_layernorm(hidden_states)
-        # Fully Connected
-        attention_output, self_cross_attn_weights, present_cross_key_value = self.cross_attn(
-            hidden_states=cross_input,
-            encoder_outputs=encoder_outputs,
-            attention_mask=cross_attention_mask,
-            past_key_value=past_key_value[-2:] if past_key_value is not None else None,
-            output_attentions=output_attentions,
-            use_cache=use_cache,
-            )
-        hidden_states = hidden_states + attention_output
-        mlp_input = self.post_attention_layernorm(hidden_states)
-        mlp_output = self.mlp(mlp_input, token_type_ids=token_type_ids)
-        hidden_states = mlp_output + hidden_states
-
-        outputs = (hidden_states,)
-
-        if output_attentions:
-            outputs += (self_attn_weights,)
-
-        if use_cache:
-            outputs += (present_key_value+present_cross_key_value,)
-
-        return outputs  # type: ignore
-
-
-class CogAgentPreTrainedModel(PreTrainedModel):
-    config_class = CogAgentConfig
-    base_model_prefix = "model"
-    supports_gradient_checkpointing = False
-    _no_split_modules = ["CogAgentDecoderLayer"]
-    _skip_keys_device_placement = "past_key_values"
-
-    def _init_weights(self, module):
-        std = self.config.initializer_range
-        if isinstance(module, nn.Linear):
-            module.weight.data.normal_(mean=0.0, std=std)
-            if module.bias is not None:
-                module.bias.data.zero_()
-        elif isinstance(module, nn.Embedding):
-            module.weight.data.normal_(mean=0.0, std=std)
-            if module.padding_idx is not None:
-                module.weight.data[module.padding_idx].zero_()
-
-
-def is_empty(images_list: Optional[List[List[torch.Tensor]]]):
-    if images_list is None or len(images_list) == 0:
-        return True
-    for image_list in images_list:
-        if len(image_list):
-            return False
-    return True
-
-
-def build_position_ids(x: "torch.BoolTensor(B, L)", attention_mask: Optional["torch.BoolTensor(B, L)"] = None) -> "torch.LongTensor(B, L)":
-    if attention_mask is not None:
-        tmp = x.clone()
-        tmp[~(attention_mask.bool())] = -1
-    else:
-        tmp = x.clone()
-    # image boi eoi token as LANGUAGE_TOKEN_TYPE
-    is_boi_eoi = torch.zeros_like(x, dtype=torch.bool)
-    is_boi_eoi[:, 1:] |= (tmp[:, 1:] == VISION_TOKEN_TYPE) & (tmp[:, :-1] == LANGUAGE_TOKEN_TYPE)
-    is_boi_eoi[:, 0] |= (tmp[:, 0] == VISION_TOKEN_TYPE)
-    is_boi_eoi[:, :-1] |= (tmp[:, :-1] == VISION_TOKEN_TYPE) & (tmp[:, 1:] == LANGUAGE_TOKEN_TYPE)
-    is_boi_eoi[:, -1] |= (tmp[:, -1] == VISION_TOKEN_TYPE)
-    tmp[is_boi_eoi] = LANGUAGE_TOKEN_TYPE
-    # final position ids
-    y = torch.zeros_like(x, dtype=torch.long)
-    y[:, 1:] = (tmp[:, 1:] == LANGUAGE_TOKEN_TYPE) | ((tmp[:, 1:] == VISION_TOKEN_TYPE) & (tmp[:, :-1] == LANGUAGE_TOKEN_TYPE))
-    y = y.cumsum(dim=-1)
-    return y
-
-
-class CogAgentModel(CogAgentPreTrainedModel):
-    def __init__(self, config):
-        super().__init__(config)
-        self.padding_idx = config.pad_token_id
-        self.vocab_size = config.vocab_size
-
-        self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
-        self.layers = nn.ModuleList([CogAgentDecoderLayer(config) for _ in range(config.num_hidden_layers)])
-        self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
-
-        self.vision = EVA2CLIPModel(config)
-        self.cross_vision = CrossVisionModel(config)
-
-        self.gradient_checkpointing = False
-        # Initialize weights and apply final processing
-        self.post_init()
-
-    def encode_images(self, images: List[List[torch.Tensor]]) -> torch.Tensor:
-        images_list, images = images, []
-
-        images = []
-        for image_list in images_list:
-            for image in image_list:
-                images.append(image)
-
-        images = torch.stack(images)
-        images_features = self.vision(images)
-        return images_features
-
-    def encode_cross_images(self, images: List[List[torch.Tensor]]) -> torch.Tensor:
-        images_list, images = images, []
-
-        images = []
-        for image_list in images_list:
-            for image in image_list:
-                images.append(image)
-
-        images = torch.stack(images)
-        encoder_outputs = self.cross_vision(images)
-        return encoder_outputs
-    
-    def forward(
-            self,
-            input_ids: torch.LongTensor = None,
-            images: List[List[torch.Tensor]] = None,
-            cross_images: List[List[torch.Tensor]] = None,
-            token_type_ids: Optional[torch.LongTensor] = None,
-            attention_mask: Optional[torch.Tensor] = None,
-            cross_attention_mask: Optional[torch.Tensor] = None,
-            position_ids: Optional[torch.LongTensor] = None,
-            past_key_values: Optional[List[torch.FloatTensor]] = None,
-            inputs_embeds: Optional[torch.FloatTensor] = None,
-            use_cache: Optional[bool] = None,
-            output_attentions: Optional[bool] = None,
-            output_hidden_states: Optional[bool] = None,
-            return_dict: Optional[bool] = None,
-    ) -> Union[Tuple, BaseModelOutputWithPast]:
-        """take care of image_encode, token_type_ids, position_ids and (attention_mask = None is fine)"""
-
-        if past_key_values is not None:
-            encoder_outputs = None
-            # generate mode with past_key_values. the image features are already mapped
-        else:
-            # not allow for inputs_embeds, because we want to process image feature
-            assert input_ids is not None and inputs_embeds is None, f"{input_ids} {inputs_embeds}"
-            if not is_empty(images):  # multi-modality
-                assert token_type_ids is not None, f"multi-modality requires `token_type_ids`!"
-                assert len(input_ids) == len(images), f"{len(input_ids)} {len(images)}"
-                inputs_embeds = self.embed_tokens(input_ids)
-                images_features = self.encode_images(images)
-                encoder_outputs = self.encode_cross_images(cross_images)
-                images_features = rearrange(images_features, 'b n d -> (b n) d')
-                images_features = images_features.to(dtype=inputs_embeds.dtype, device=inputs_embeds.device)
-                inputs_embeds = inputs_embeds.index_put([token_type_ids == VISION_TOKEN_TYPE], images_features)
-            else:  # single-modality
-                if token_type_ids is None:
-                    token_type_ids = torch.ones_like(input_ids, dtype=torch.long, device=input_ids.device) * LANGUAGE_TOKEN_TYPE
-                assert not (token_type_ids == VISION_TOKEN_TYPE).any(), f"{(token_type_ids == VISION_TOKEN_TYPE).sum()}"
-                inputs_embeds = self.embed_tokens(input_ids)
-                encoder_outputs = None
-
-            if position_ids is None:
-                position_ids = build_position_ids(token_type_ids, attention_mask)
-            input_ids = None
-
-        return self.llm_forward(
-            input_ids=input_ids,
-            encoder_outputs=encoder_outputs,
-            token_type_ids=token_type_ids,
-            attention_mask=attention_mask,
-            cross_attention_mask=cross_attention_mask,
-            position_ids=position_ids,
-            past_key_values=past_key_values,
-            inputs_embeds=inputs_embeds,
-            use_cache=use_cache,
-            output_attentions=output_attentions,
-            output_hidden_states=output_hidden_states,
-            return_dict=return_dict,
-        )
-
-    def llm_forward(
-            self,
-            input_ids: torch.LongTensor = None,
-            encoder_outputs: torch.LongTensor = None,
-            token_type_ids: torch.LongTensor = None,
-            attention_mask: Optional[torch.Tensor] = None,
-            cross_attention_mask: Optional[torch.Tensor] = None,
-            position_ids: Optional[torch.LongTensor] = None,
-            past_key_values: Optional[List[torch.FloatTensor]] = None,
-            inputs_embeds: Optional[torch.FloatTensor] = None,
-            use_cache: Optional[bool] = None,
-            output_attentions: Optional[bool] = None,
-            output_hidden_states: Optional[bool] = None,
-            return_dict: Optional[bool] = None,
-    ) -> Union[Tuple, BaseModelOutputWithPast]:
-        """largely copy from llama forward and adapt for CogAgent with `token_type_ids`"""
-        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
-        output_hidden_states = (
-            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
-        )
-        use_cache = use_cache if use_cache is not None else self.config.use_cache
-
-        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
-
-        # retrieve input_ids and inputs_embeds
-        if input_ids is not None and inputs_embeds is not None:
-            raise ValueError("You cannot specify both decoder_input_ids and decoder_inputs_embeds at the same time")
-        elif input_ids is not None:
-            batch_size, seq_length = input_ids.shape
-        elif inputs_embeds is not None:
-            batch_size, seq_length, _ = inputs_embeds.shape
-        else:
-            raise ValueError("You have to specify either decoder_input_ids or decoder_inputs_embeds")
-
-        seq_length_with_past = seq_length
-        past_key_values_length = 0
-
-        if past_key_values is not None:
-            past_key_values_length = past_key_values[0][0].shape[2]
-            seq_length_with_past = seq_length_with_past + past_key_values_length
-
-        if position_ids is None:
-            device = input_ids.device if input_ids is not None else inputs_embeds.device
-            position_ids = torch.arange(
-                past_key_values_length, seq_length + past_key_values_length, dtype=torch.long, device=device
-            )
-            position_ids = position_ids.unsqueeze(0).view(-1, seq_length)
-        else:
-            position_ids = position_ids.view(-1, seq_length).long()
-
-        if inputs_embeds is None:
-            inputs_embeds = self.embed_tokens(input_ids)
-        # embed positions
-        if attention_mask is None:
-            attention_mask = torch.ones(
-                (batch_size, seq_length_with_past), dtype=torch.bool, device=inputs_embeds.device
-            )
-        if cross_attention_mask is None:
-            cross_attention_mask = torch.ones(
-                (batch_size, 1), dtype=torch.bool, device=inputs_embeds.device
-            )
-        attention_mask = self._prepare_decoder_attention_mask(
-            attention_mask, (batch_size, seq_length), inputs_embeds, past_key_values_length
-        )
-
-        hidden_states = inputs_embeds
-
-        # decoder layers
-        all_hidden_states = () if output_hidden_states else None
-        all_self_attns = () if output_attentions else None
-        next_decoder_cache = () if use_cache else None
-
-        for idx, decoder_layer in enumerate(self.layers):
-            if output_hidden_states:
-                all_hidden_states += (hidden_states,)
-
-            past_key_value = past_key_values[idx] if past_key_values is not None else None
-            layer_outputs = decoder_layer(
-                hidden_states,
-                encoder_outputs=encoder_outputs,
-                token_type_ids=token_type_ids,
-                attention_mask=attention_mask,
-                cross_attention_mask=cross_attention_mask,
-                position_ids=position_ids,
-                past_key_value=past_key_value,
-                output_attentions=output_attentions,
-                use_cache=use_cache,
-            )
-            hidden_states = layer_outputs[0]
-
-            if use_cache:
-                next_decoder_cache += (layer_outputs[2 if output_attentions else 1],)
-
-            if output_attentions:
-                all_self_attns += (layer_outputs[1],)
-
-        hidden_states = self.norm(hidden_states)
-
-        # add hidden states from the last decoder layer
-        if output_hidden_states:
-            all_hidden_states += (hidden_states,)
-
-        next_cache = next_decoder_cache if use_cache else None
-        if not return_dict:
-            return tuple(v for v in [hidden_states, next_cache, all_hidden_states, all_self_attns] if v is not None)
-        return BaseModelOutputWithPast(
-            last_hidden_state=hidden_states,
-            past_key_values=next_cache,
-            hidden_states=all_hidden_states,
-            attentions=all_self_attns,
-        )
-
-    def get_input_embeddings(self):
-        return self.embed_tokens
-
-    def set_input_embeddings(self, value):
-        self.embed_tokens = value
-
-    # noinspection PyMethodMayBeStatic
-    # Copied from transformers.models.bart.modeling_bart.BartDecoder._prepare_decoder_attention_mask
-    def _prepare_decoder_attention_mask(self, attention_mask, input_shape, inputs_embeds, past_key_values_length):
-        # create causal mask
-        # [bsz, seq_len] -> [bsz, 1, tgt_seq_len, src_seq_len]
-        combined_attention_mask = None
-        if input_shape[-1] > 1:
-            combined_attention_mask = _make_causal_mask(
-                input_shape,
-                inputs_embeds.dtype,
-                device=inputs_embeds.device,
-                past_key_values_length=past_key_values_length,
-            )
-
-        if attention_mask is not None:
-            # [bsz, seq_len] -> [bsz, 1, tgt_seq_len, src_seq_len]
-            expanded_attn_mask = _expand_mask(attention_mask, inputs_embeds.dtype, tgt_len=input_shape[-1]).to(
-                inputs_embeds.device
-            )
-            combined_attention_mask = (
-                expanded_attn_mask if combined_attention_mask is None else expanded_attn_mask + combined_attention_mask
-            )
-
-        return combined_attention_mask
-
-def chat_old_history_to_prompt(history, query):
-    prompt = "<EOI>Question: "
-    for i, (old_query, response) in enumerate(history):
-        prompt += old_query + " Answer: " + response + "\nQuestion: "
-    prompt += query + " Answer:"
-    return prompt
-
-def chat_history_to_prompt(history, query):
-    prompt = " [INST] "
-    for i, (old_query, response) in enumerate(history):
-        prompt += old_query + " [/INST] " + response + " [INST] "
-    prompt += query + " [/INST] "
-    return prompt
-
-
-def base_history_to_prompt(history, query):
-    prompt = query
-    return prompt
-
-
-_history_to_prompt = {
-    "base": base_history_to_prompt,
-    "chat": chat_history_to_prompt,
-    "chat_old": chat_old_history_to_prompt
-}
-
-
-class CogAgentForCausalLM(CogAgentPreTrainedModel):
-    _auto_class = "AutoModelForCausalLM"
-
-    def __init__(self, config):
-        super().__init__(config)
-        self.model = CogAgentModel(config)
-        self.vocab_size = config.vocab_size
-        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
-
-        # Initialize weights and apply final processing
-        self.post_init()
-
-    def get_input_embeddings(self):
-        return self.model.embed_tokens
-
-    def set_input_embeddings(self, value):
-        self.model.embed_tokens = value
-
-    def get_output_embeddings(self):
-        return self.lm_head
-
-    def set_output_embeddings(self, new_embeddings):
-        self.lm_head = new_embeddings
-
-    def set_decoder(self, decoder):
-        self.model = decoder
-
-    def get_decoder(self):
-        return self.model
-
-    def forward(
-            self,
-            input_ids: torch.LongTensor = None,
-            images: List[List[torch.Tensor]] = None,
-            cross_images: List[List[torch.Tensor]] = None,
-            token_type_ids: Optional[torch.LongTensor] = None,
-            attention_mask: Optional[torch.Tensor] = None,
-            position_ids: Optional[torch.LongTensor] = None,
-            past_key_values: Optional[List[torch.FloatTensor]] = None,
-            inputs_embeds: Optional[torch.FloatTensor] = None,
-            use_cache: Optional[bool] = None,
-            output_attentions: Optional[bool] = None,
-            output_hidden_states: Optional[bool] = None,
-            return_dict: Optional[bool] = None,
-            labels: Optional[torch.LongTensor] = None,
-    ) -> Union[Tuple, CausalLMOutputWithPast]:
-        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
-        output_hidden_states = (
-            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
-        )
-        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
-
-        # decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)
-        outputs = self.model(
-            input_ids=input_ids,
-            images=images,
-            cross_images=cross_images,
-            token_type_ids=token_type_ids,
-            attention_mask=attention_mask,
-            position_ids=position_ids,
-            past_key_values=past_key_values,
-            inputs_embeds=inputs_embeds,
-            use_cache=use_cache,
-            output_attentions=output_attentions,
-            output_hidden_states=output_hidden_states,
-            return_dict=return_dict,
-        )
-
-        hidden_states = outputs[0]
-        logits = self.lm_head(hidden_states)
-        logits = logits.float()
-
-        loss = None
-        if labels is not None:
-            # Shift so that tokens < n predict n
-            shift_logits = logits[..., :-1, :].contiguous()
-            shift_labels = labels[..., 1:].contiguous()
-            # Flatten the tokens
-            loss_fct = CrossEntropyLoss()
-            shift_logits = shift_logits.view(-1, self.config.vocab_size)
-            shift_labels = shift_labels.view(-1)
-            # Enable model parallelism
-            shift_labels = shift_labels.to(shift_logits.device)
-            loss = loss_fct(shift_logits, shift_labels)
-
-        if not return_dict:
-            output = (logits,) + outputs[1:]
-            return (loss,) + output if loss is not None else output
-
-        return CausalLMOutputWithPast(
-            loss=loss,
-            logits=logits,
-            past_key_values=outputs.past_key_values,
-            hidden_states=outputs.hidden_states,
-            attentions=outputs.attentions,
-        )
-
-    def _prepare_attention_mask_for_generation(
-            self,
-            inputs: torch.Tensor,
-            pad_token_id: Optional[int],
-            eos_token_id: Optional[Union[int, List[int]]],
-    ) -> torch.LongTensor:
-        return torch.ones(inputs.shape[:2], dtype=torch.long, device=inputs.device)  # type: ignore
-
-    def prepare_inputs_for_generation(
-            self, input_ids, token_type_ids, images=None, cross_images=None, past_key_values=None, attention_mask=None, inputs_embeds=None, **kwargs
-    ):
-        # build position_ids if needed
-        position_ids = kwargs.get("position_ids", None)
-        if position_ids is None:
-            position_ids = build_position_ids(token_type_ids, attention_mask)
-
-        if past_key_values:
-            input_ids = input_ids[:, -1:]
-            token_type_ids = token_type_ids[:, -1:]
-            position_ids = position_ids[:, -1:]
-
-        # if `inputs_embeds` are passed, we only want to use them in the 1st generation step
-        if inputs_embeds is not None and past_key_values is None:
-            model_inputs = {"inputs_embeds": inputs_embeds}
-        else:
-            model_inputs = {"input_ids": input_ids}
-
-        model_inputs.update(
-            {
-                "token_type_ids": token_type_ids,
-                "images": images,
-                "cross_images": cross_images,
-                "position_ids": position_ids,
-                "past_key_values": past_key_values,
-                "use_cache": kwargs.get("use_cache"),
-                "attention_mask": attention_mask,
-            }
-        )
-        return model_inputs
-
-    def _update_model_kwargs_for_generation(
-            self,
-            outputs: "ModelOutput",
-            model_kwargs: Dict[str, Any],
-            is_encoder_decoder: bool = False,
-            standardize_cache_format: bool = False,
-    ) -> Dict[str, Any]:
-        # update past_key_values
-        model_kwargs["past_key_values"] = self._extract_past_from_model_output(
-            outputs, standardize_cache_format=standardize_cache_format
-        )
-        if getattr(outputs, "state", None) is not None:
-            model_kwargs["state"] = outputs.state
-
-        # update token_type_ids with last value
-        if "token_type_ids" in model_kwargs:
-            token_type_ids = model_kwargs["token_type_ids"]
-            new_token_type_ids = torch.ones(size=(token_type_ids.shape[0], 1), dtype=token_type_ids.dtype, device=token_type_ids.device) * LANGUAGE_TOKEN_TYPE
-            model_kwargs["token_type_ids"] = torch.cat([token_type_ids, new_token_type_ids], dim=-1)
-
-        if not is_encoder_decoder:
-            # update attention mask
-            if "attention_mask" in model_kwargs:
-                attention_mask = model_kwargs["attention_mask"]
-                model_kwargs["attention_mask"] = torch.cat(
-                    [attention_mask, attention_mask.new_ones((attention_mask.shape[0], 1))], dim=-1
-                )
-        else:
-            # update decoder attention mask
-            if "decoder_attention_mask" in model_kwargs:
-                decoder_attention_mask = model_kwargs["decoder_attention_mask"]
-                model_kwargs["decoder_attention_mask"] = torch.cat(
-                    [decoder_attention_mask, decoder_attention_mask.new_ones((decoder_attention_mask.shape[0], 1))],
-                    dim=-1,
-                )
-
-        return model_kwargs
-
-    def _reorder_cache(self, past_key_values, beam_idx):
-        reordered_past = ()
-        for layer_past in past_key_values:
-            reordered_past += (
-                tuple(past_state.index_select(0, beam_idx.to(past_state.device)) for past_state in layer_past),
-            )
-        return reordered_past
-
-    def build_conversation_input_ids(
-            self,
-            tokenizer: "PreTrainedTokenizer",
-            *,
-            query: str,
-            history: Optional[List[Tuple[str, str]]] = None,
-            images: Optional[List["PIL.Image"]] = None,
-            template_version: Optional[Literal["base", "chat", "vqa"]] = None,
-    ):
-        image_size: int = self.config.vision_config['image_size']
-        cross_image_size: int = self.config.cross_image_size
-        patch_size: int = self.config.vision_config['patch_size']
-        template_version = template_version or self.config.template_version
-        assert images is None or len(images) <= 1, f"not support multi images by now."
-        history = history or []
-        text = _history_to_prompt[template_version](history, query)
-
-        input_ids = [tokenizer.bos_token_id]
-        token_type_ids = [LANGUAGE_TOKEN_TYPE]
-        if images is not None and len(images) == 1:
-            ori = images
-            # vision
-            transform = transforms.Compose(
-                [
-                    transforms.Resize(
-                        (image_size, image_size), interpolation=transforms.InterpolationMode.BICUBIC
-                    ),
-                    transforms.ToTensor(),
-                    transforms.Normalize((0.48145466, 0.4578275, 0.40821073), (0.26862954, 0.26130258, 0.27577711)),
-                ]
-            )
-            images = [transform(ori[0])]
-            cross_transform = transforms.Compose(
-                [
-                    transforms.Resize(
-                        (cross_image_size, cross_image_size), interpolation=transforms.InterpolationMode.BICUBIC
-                    ),
-                    transforms.ToTensor(),
-                    transforms.Normalize((0.48145466, 0.4578275, 0.40821073), (0.26862954, 0.26130258, 0.27577711)),
-                ]
-            )
-            cross_images = [cross_transform(ori[0])]
-            # language
-            vision_token_num = (image_size // patch_size) * (image_size // patch_size) + 2
-            input_ids += [tokenizer.pad_token_id] * vision_token_num
-            token_type_ids += [VISION_TOKEN_TYPE] * vision_token_num
-        text_ids = tokenizer.encode(text, add_special_tokens=False)
-
-        input_ids += text_ids
-        token_type_ids += [LANGUAGE_TOKEN_TYPE] * len(text_ids)
-        attention_mask = [1] * len(input_ids)
-
-        return {
-            'input_ids': torch.tensor(input_ids, dtype=torch.long),
-            'token_type_ids': torch.tensor(token_type_ids, dtype=torch.long),
-            'attention_mask': torch.tensor(attention_mask, dtype=torch.long),
-            'images': images,
-            'cross_images': cross_images
-        }
+"""largely copy from llama and adapt for CogAgent"""
+import warnings
+from typing import TYPE_CHECKING, Optional, Tuple, List, Union, Literal, Dict, Any
+
+import math
+import torch
+from torch import nn
+from torch.nn import CrossEntropyLoss
+from torchvision import transforms
+from einops import rearrange
+
+from transformers import PreTrainedModel, PreTrainedTokenizer
+from transformers.utils.logging import get_logger
+from transformers.activations import ACT2FN
+from transformers.modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast
+
+from .configuration_cogagent import CogAgentConfig
+# from .util import FastRotaryEmbedding
+from torch.nn import functional as F
+from .visual import EVA2CLIPModel
+from .cross_visual import CrossVisionModel
+
+if TYPE_CHECKING:
+    from transformers.utils import ModelOutput
+
+logger = get_logger(__name__)
+
+LANGUAGE_TOKEN_TYPE = 0
+VISION_TOKEN_TYPE = 1
+
+
+# Copied from transformers.models.bart.modeling_bart._make_causal_mask
+def _make_causal_mask(
+        input_ids_shape: torch.Size, dtype: torch.dtype, device: torch.device, past_key_values_length: int = 0
+):
+    """
+    Make causal mask used for bi-directional self-attention.
+    """
+    bsz, tgt_len = input_ids_shape
+    mask = torch.full((tgt_len, tgt_len), torch.finfo(dtype).min, device=device)
+    mask_cond = torch.arange(mask.size(-1), device=device)
+    mask.masked_fill_(mask_cond < (mask_cond + 1).view(mask.size(-1), 1), 0)
+    mask = mask.to(dtype)
+
+    if past_key_values_length > 0:
+        mask = torch.cat([torch.zeros(tgt_len, past_key_values_length, dtype=dtype, device=device), mask], dim=-1)
+    return mask[None, None, :, :].expand(bsz, 1, tgt_len, tgt_len + past_key_values_length)
+
+
+# Copied from transformers.models.bart.modeling_bart._expand_mask
+def _expand_mask(mask: torch.Tensor, dtype: torch.dtype, tgt_len: Optional[int] = None):
+    """
+    Expands attention_mask from `[bsz, seq_len]` to `[bsz, 1, tgt_seq_len, src_seq_len]`.
+    """
+    bsz, src_len = mask.size()
+    tgt_len = tgt_len if tgt_len is not None else src_len
+
+    expanded_mask = mask[:, None, None, :].expand(bsz, 1, tgt_len, src_len).to(dtype)
+
+    inverted_mask = 1.0 - expanded_mask
+
+    return inverted_mask.masked_fill(inverted_mask.to(torch.bool), torch.finfo(dtype).min)
+
+
+class RMSNorm(nn.Module):
+    def __init__(self, hidden_size, eps=1e-6):
+        super().__init__()
+        self.weight = nn.Parameter(torch.ones(hidden_size))
+        self.variance_epsilon = eps
+
+    def forward(self, hidden_states):
+        input_dtype = hidden_states.dtype
+        hidden_states = hidden_states.to(torch.float32)
+        variance = hidden_states.pow(2).mean(-1, keepdim=True)
+        hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
+        return (self.weight * hidden_states).to(input_dtype)
+
+
+class MLP(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.hidden_size = config.hidden_size
+        self.intermediate_size = config.intermediate_size
+        self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
+        self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
+        self.down_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=False)
+        self.act_fn = ACT2FN[config.hidden_act]
+
+    def forward(self, x):
+        down_proj = self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))
+        return down_proj
+
+
+def get_expert_mask(token_type_ids: "torch.LongTensor(B, L)") -> "[torch.BoolTensor(B, L), torch.BoolTensor(B, L)]":
+    vision_token_mask = torch.zeros_like(token_type_ids, dtype=torch.bool)
+    vision_token_mask[:, :-1] = (token_type_ids[:, :-1] == VISION_TOKEN_TYPE) & (token_type_ids[:, 1:] == VISION_TOKEN_TYPE)
+    language_token_mask = ~vision_token_mask
+    return vision_token_mask, language_token_mask
+
+
+class VisionExpertMLP(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.language_mlp = MLP(config)
+        self.vision_mlp = MLP(config)
+
+    def forward(self, hidden_states: "torch.Tensor(B, L, D)", token_type_ids: "torch.LongTensor(B, L)"):
+        output = torch.empty(hidden_states.shape, dtype=hidden_states.dtype, device=hidden_states.device)
+        vision_token_mask, language_token_mask = get_expert_mask(token_type_ids)
+        output[vision_token_mask] = self.vision_mlp(hidden_states[vision_token_mask])
+        output[language_token_mask] = self.language_mlp(hidden_states[language_token_mask])
+        return output
+
+
+def attention_fn(
+        query_layer: "torch.tensor(B, H, L, HD)",
+        key_layer: "torch.tensor(B, H, L, HD)",
+        value_layer: "torch.tensor(B, H, L, HD)",
+        attention_mask: "torch.tensor(B, H, L, HD)",
+        *,
+        scaling_attention_score: bool = True,
+        attention_dropout: nn.Module = None
+):
+    attention_mask_bool = (attention_mask == 0)
+    is_low_triangle = (attention_mask_bool == torch.ones_like(attention_mask_bool, dtype=torch.float).tril()).all()
+    is_full = (attention_mask_bool > 0).all()
+    if not (int(torch.__version__.split('.')[0]) >= 2):
+        warnings.warn("It's recommended to use torch2.0 or higher.")
+    if int(torch.__version__.split('.')[0]) >= 2 and scaling_attention_score and (is_full or is_low_triangle):
+        dropout_p = 0. if attention_dropout is None or not attention_dropout.training else attention_dropout.p
+        return torch.nn.functional.scaled_dot_product_attention(
+            query_layer, key_layer, value_layer,
+            attn_mask=None,
+            dropout_p=dropout_p,
+            is_causal=not is_full
+        )
+    else:
+        if scaling_attention_score:
+            query_layer = query_layer / math.sqrt(query_layer.shape[-1])
+        attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2))
+        attention_scores = attention_scores + attention_mask
+        attention_scores = nn.functional.softmax(attention_scores, dim=-1, dtype=torch.float32).to(query_layer.dtype)
+        if attention_dropout is not None:
+            attention_scores = attention_dropout(attention_scores)
+        context_layer = torch.matmul(attention_scores, value_layer)
+        return context_layer
+
+class RotaryEmbedding(torch.nn.Module):
+    def __init__(self, dim, max_position_embeddings=2048, base=10000, device=None):
+        super().__init__()
+
+        self.dim = dim
+        self.max_position_embeddings = max_position_embeddings
+        self.base = base
+        inv_freq = self._compute_inv_freq(device)
+        self.register_buffer("inv_freq", inv_freq)
+        self.max_seq_len_cached = 0
+
+    def _compute_inv_freq(self, device=None):
+        return 1.0 / (
+                self.base
+                ** (torch.arange(0, self.dim, 2, device=device) / self.dim)
+        )
+
+    def _set_cos_sin_cache(self, seq_len, device, dtype):
+        self.max_seq_len_cached = seq_len
+        t = torch.arange(self.max_seq_len_cached, device=device, dtype=self.inv_freq.dtype)
+
+        freqs = torch.einsum("i,j->ij", t, self.inv_freq)
+        # Different from paper, but it uses a different permutation in order to obtain the same calculation
+        emb = torch.cat((freqs, freqs), dim=-1)
+        self.register_buffer("cos_cached", emb.cos()[:, None, :].to(dtype), persistent=False)
+        self.register_buffer("sin_cached", emb.sin()[:, None, :].to(dtype), persistent=False)
+
+    def forward(self, x, seq_len):
+        # x: [bs, num_attention_heads, seq_len, head_size]
+        if seq_len > self.max_seq_len_cached:
+            self._set_cos_sin_cache(seq_len=seq_len, device=x.device, dtype=x.dtype)
+
+        return (
+            self.cos_cached[:seq_len, ...].to(dtype=x.dtype),
+            self.sin_cached[:seq_len, ...].to(dtype=x.dtype),
+        )
+
+
+def rotate_half(x):
+    x1, x2 = x[..., :x.shape[-1] // 2], x[..., x.shape[-1] // 2:]
+    return torch.cat((-x2, x1), dim=x1.ndim - 1)
+
+
+def apply_rotary_pos_emb_index_bhs(q, k, cos, sin, position_id):
+    # batch_size, num_head, seq_len, hidden_size
+    cos, sin = F.embedding(position_id, cos.squeeze(1)).unsqueeze(1), \
+        F.embedding(position_id, sin.squeeze(1)).unsqueeze(1)
+    q, k = (q * cos) + (rotate_half(q) * sin), (k * cos) + (rotate_half(k) * sin)
+    return q, k
+
+class VisionExpertAttention(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.config = config
+        self.hidden_size = config.hidden_size
+        self.num_heads = config.num_attention_heads
+        self.head_dim = self.hidden_size // self.num_heads
+        self.max_position_embeddings = config.max_position_embeddings
+
+        self.rotary_emb = RotaryEmbedding(self.head_dim)
+        self.vision_expert_query_key_value = nn.Linear(self.hidden_size, self.hidden_size * 3, bias=False)
+        self.vision_expert_dense = nn.Linear(self.hidden_size, self.hidden_size, bias=False)
+        self.language_expert_query_key_value = nn.Linear(self.hidden_size, self.hidden_size * 3, bias=False)
+        self.language_expert_dense = nn.Linear(self.hidden_size, self.hidden_size, bias=False)
+
+    def _transpose_for_scores(self, tensor):
+        """Transpose a 3D tensor [B, L, H*HD] into a 4D tensor with size [B H L HD]."""
+        new_tensor_shape = tensor.size()[:-1] + (self.num_heads, self.head_dim)
+        tensor = tensor.view(*new_tensor_shape)
+        return tensor.permute(0, 2, 1, 3)
+
+    def forward(
+            self,
+            hidden_states: torch.Tensor,
+            token_type_ids: torch.LongTensor,
+            position_ids: torch.LongTensor,
+            attention_mask: Optional[torch.Tensor] = None,
+            past_key_value: Optional[Tuple[torch.Tensor]] = None,
+            output_attentions: bool = False,
+            use_cache: bool = False,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+        bsz, q_len, _ = hidden_states.size()
+        vision_token_mask, language_token_mask = get_expert_mask(token_type_ids)
+
+        shape = list(hidden_states.shape)
+        shape[-1] = shape[-1] * 3
+        mixed_raw_layer = torch.empty(shape, dtype=hidden_states.dtype, device=hidden_states.device)
+        mixed_raw_layer[vision_token_mask] = self.vision_expert_query_key_value(hidden_states[vision_token_mask])
+        mixed_raw_layer[language_token_mask] = self.language_expert_query_key_value(hidden_states[language_token_mask])
+
+        query_states, key_states, value_states = torch.split(mixed_raw_layer, self.hidden_size, dim=-1)
+        query_states = self._transpose_for_scores(query_states)  # B, H, L, HD
+        key_states = self._transpose_for_scores(key_states)  # B, H, L, HD
+        value_states = self._transpose_for_scores(value_states)  # B, H, L, HD
+
+        kv_seq_len = key_states.shape[-2]
+        if past_key_value is not None:
+            kv_seq_len += past_key_value[0].shape[-2]
+
+        cos, sin = self.rotary_emb(value_states, seq_len=position_ids.max() + 1)
+        query_states, key_states = apply_rotary_pos_emb_index_bhs(query_states, key_states, cos, sin, position_ids)
+
+        if past_key_value is not None:
+            key_states = torch.cat([past_key_value[0], key_states], dim=2)
+            value_states = torch.cat([past_key_value[1], value_states], dim=2)
+
+        past_key_value = (key_states, value_states) if use_cache else None
+
+        context_layer = attention_fn(
+            query_layer=query_states, key_layer=key_states, value_layer=value_states, attention_mask=attention_mask,
+            scaling_attention_score=True, attention_dropout=None)
+        if context_layer.size() != (bsz, self.num_heads, q_len, self.head_dim):
+            raise ValueError(
+                f"`attn_output` should be of size {(bsz, self.num_heads, q_len, self.head_dim)}, but is"
+                f" {context_layer.size()}"
+            )
+        context_layer = context_layer.transpose(1, 2).contiguous().reshape(bsz, q_len, self.hidden_size)
+
+        attn_output = torch.empty(context_layer.shape, dtype=hidden_states.dtype, device=hidden_states.device)
+        attn_output[vision_token_mask] = self.vision_expert_dense(context_layer[vision_token_mask])
+        attn_output[language_token_mask] = self.language_expert_dense(context_layer[language_token_mask])
+
+        if output_attentions:
+            warnings.warn("output_attentions is not implemented.")
+
+        return attn_output, None, past_key_value
+
+class CrossAttention(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.config = config
+        self.hidden_size = config.hidden_size
+        self.cross_hidden_size = config.cross_hidden_size
+        self.cross_compute_hidden_size = config.cross_compute_hidden_size
+        self.num_heads = config.num_attention_heads
+        self.head_dim = self.hidden_size // self.num_heads
+        self.cross_head_dim = self.cross_compute_hidden_size // self.num_heads
+        self.max_position_embeddings = config.max_position_embeddings
+
+        self.query = nn.Linear(self.hidden_size, self.cross_compute_hidden_size, bias=False)
+        self.key_value = nn.Linear(self.cross_hidden_size, self.cross_compute_hidden_size * 2, bias=False)
+        self.dense = nn.Linear(self.cross_compute_hidden_size, self.hidden_size, bias=False)
+
+    def _transpose_for_scores(self, tensor):
+        """Transpose a 3D tensor [B, L, H*HD] into a 4D tensor with size [B H L HD]."""
+        new_tensor_shape = tensor.size()[:-1] + (self.num_heads, self.cross_head_dim)
+        tensor = tensor.view(*new_tensor_shape)
+        return tensor.permute(0, 2, 1, 3)
+
+    def forward(
+            self,
+            hidden_states: torch.Tensor,
+            encoder_outputs: torch.LongTensor,
+            attention_mask: Optional[torch.Tensor] = None,
+            past_key_value: Optional[Tuple[torch.Tensor]] = None,
+            output_attentions: bool = False,
+            use_cache: bool = False,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+        bsz, q_len, _ = hidden_states.size()
+
+        shape = list(hidden_states.shape)
+        shape[-1] = shape[-1] * 3
+
+        mixed_query_layer = self.query(hidden_states)
+        if past_key_value is None:
+            mixed_x_layer = self.key_value(encoder_outputs)
+            mixed_key_layer, mixed_value_layer = torch.split(mixed_x_layer, self.cross_compute_hidden_size, dim=-1)
+            key_states = self._transpose_for_scores(mixed_key_layer)  # B, H, L, HD
+            value_states = self._transpose_for_scores(mixed_value_layer)  # B, H, L, HD
+        else:
+            key_states, value_states = past_key_value
+
+        query_states = self._transpose_for_scores(mixed_query_layer)  # B, H, L, HD
+
+        past_key_value = (key_states, value_states) if use_cache else None
+
+        context_layer = attention_fn(
+            query_layer=query_states, key_layer=key_states, value_layer=value_states, attention_mask=attention_mask,
+            scaling_attention_score=True, attention_dropout=None)
+        if context_layer.size() != (bsz, self.num_heads, q_len, self.cross_head_dim):
+            raise ValueError(
+                f"`cross_attn_output` should be of size {(bsz, self.num_heads, q_len, self.cross_head_dim)}, but is"
+                f" {context_layer.size()}"
+            )
+        context_layer = context_layer.transpose(1, 2).contiguous().reshape(bsz, q_len, self.cross_hidden_size)
+
+        attn_output = self.dense(context_layer)
+
+        if output_attentions:
+            warnings.warn("output_attentions is not implemented.")
+
+        return attn_output, None, past_key_value
+
+class CogAgentDecoderLayer(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.hidden_size = config.hidden_size
+        self.self_attn = VisionExpertAttention(config=config)
+        self.cross_attn = CrossAttention(config=config)
+        self.mlp = VisionExpertMLP(config)
+        self.input_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        self.post_attention_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        self.post_cross_attention_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+
+    def forward(
+            self,
+            hidden_states: torch.Tensor,
+            encoder_outputs: torch.Tensor,
+            token_type_ids: torch.LongTensor,
+            position_ids: torch.LongTensor,
+            attention_mask: Optional[torch.Tensor] = None,
+            cross_attention_mask: Optional[torch.Tensor] = None,
+            past_key_value: Optional[Tuple[torch.Tensor]] = None,
+            output_attentions: Optional[bool] = False,
+            use_cache: Optional[bool] = False,
+    ) -> Tuple[torch.FloatTensor, Optional[Tuple[torch.FloatTensor, torch.FloatTensor]]]:
+        residual = hidden_states
+
+        hidden_states = self.input_layernorm(hidden_states)
+
+        # Self Attention
+        hidden_states, self_attn_weights, present_key_value = self.self_attn(
+            hidden_states=hidden_states,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            attention_mask=attention_mask,
+            past_key_value=past_key_value[:2] if past_key_value is not None else None,
+            output_attentions=output_attentions,
+            use_cache=use_cache,
+        )
+        hidden_states = residual + hidden_states
+
+        cross_input = self.post_cross_attention_layernorm(hidden_states)
+        # Fully Connected
+        attention_output, self_cross_attn_weights, present_cross_key_value = self.cross_attn(
+            hidden_states=cross_input,
+            encoder_outputs=encoder_outputs,
+            attention_mask=cross_attention_mask,
+            past_key_value=past_key_value[-2:] if past_key_value is not None else None,
+            output_attentions=output_attentions,
+            use_cache=use_cache,
+            )
+        hidden_states = hidden_states + attention_output
+        mlp_input = self.post_attention_layernorm(hidden_states)
+        mlp_output = self.mlp(mlp_input, token_type_ids=token_type_ids)
+        hidden_states = mlp_output + hidden_states
+
+        outputs = (hidden_states,)
+
+        if output_attentions:
+            outputs += (self_attn_weights,)
+
+        if use_cache:
+            outputs += (present_key_value+present_cross_key_value,)
+
+        return outputs  # type: ignore
+
+
+class CogAgentPreTrainedModel(PreTrainedModel):
+    config_class = CogAgentConfig
+    base_model_prefix = "model"
+    supports_gradient_checkpointing = False
+    _no_split_modules = ["CogAgentDecoderLayer"]
+    _skip_keys_device_placement = "past_key_values"
+
+    def _init_weights(self, module):
+        std = self.config.initializer_range
+        if isinstance(module, nn.Linear):
+            module.weight.data.normal_(mean=0.0, std=std)
+            if module.bias is not None:
+                module.bias.data.zero_()
+        elif isinstance(module, nn.Embedding):
+            module.weight.data.normal_(mean=0.0, std=std)
+            if module.padding_idx is not None:
+                module.weight.data[module.padding_idx].zero_()
+
+
+def is_empty(images_list: Optional[List[List[torch.Tensor]]]):
+    if images_list is None or len(images_list) == 0:
+        return True
+    for image_list in images_list:
+        if len(image_list):
+            return False
+    return True
+
+
+def build_position_ids(x: "torch.BoolTensor(B, L)", attention_mask: Optional["torch.BoolTensor(B, L)"] = None) -> "torch.LongTensor(B, L)":
+    if attention_mask is not None:
+        tmp = x.clone()
+        tmp[~(attention_mask.bool())] = -1
+    else:
+        tmp = x.clone()
+    # image boi eoi token as LANGUAGE_TOKEN_TYPE
+    is_boi_eoi = torch.zeros_like(x, dtype=torch.bool)
+    is_boi_eoi[:, 1:] |= (tmp[:, 1:] == VISION_TOKEN_TYPE) & (tmp[:, :-1] == LANGUAGE_TOKEN_TYPE)
+    is_boi_eoi[:, 0] |= (tmp[:, 0] == VISION_TOKEN_TYPE)
+    is_boi_eoi[:, :-1] |= (tmp[:, :-1] == VISION_TOKEN_TYPE) & (tmp[:, 1:] == LANGUAGE_TOKEN_TYPE)
+    is_boi_eoi[:, -1] |= (tmp[:, -1] == VISION_TOKEN_TYPE)
+    tmp[is_boi_eoi] = LANGUAGE_TOKEN_TYPE
+    # final position ids
+    y = torch.zeros_like(x, dtype=torch.long)
+    y[:, 1:] = (tmp[:, 1:] == LANGUAGE_TOKEN_TYPE) | ((tmp[:, 1:] == VISION_TOKEN_TYPE) & (tmp[:, :-1] == LANGUAGE_TOKEN_TYPE))
+    y = y.cumsum(dim=-1)
+    return y
+
+
+class CogAgentModel(CogAgentPreTrainedModel):
+    def __init__(self, config):
+        super().__init__(config)
+        self.padding_idx = config.pad_token_id
+        self.vocab_size = config.vocab_size
+
+        self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
+        self.layers = nn.ModuleList([CogAgentDecoderLayer(config) for _ in range(config.num_hidden_layers)])
+        self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+
+        self.vision = EVA2CLIPModel(config)
+        self.cross_vision = CrossVisionModel(config)
+
+        self.gradient_checkpointing = False
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def encode_images(self, images: List[List[torch.Tensor]]) -> torch.Tensor:
+        images_list, images = images, []
+
+        images = []
+        for image_list in images_list:
+            for image in image_list:
+                images.append(image)
+
+        images = torch.stack(images)
+        images_features = self.vision(images)
+        return images_features
+
+    def encode_cross_images(self, images: List[List[torch.Tensor]]) -> torch.Tensor:
+        images_list, images = images, []
+
+        images = []
+        for image_list in images_list:
+            for image in image_list:
+                images.append(image)
+
+        images = torch.stack(images)
+        encoder_outputs = self.cross_vision(images)
+        return encoder_outputs
+    
+    def forward(
+            self,
+            input_ids: torch.LongTensor = None,
+            images: List[List[torch.Tensor]] = None,
+            cross_images: List[List[torch.Tensor]] = None,
+            token_type_ids: Optional[torch.LongTensor] = None,
+            attention_mask: Optional[torch.Tensor] = None,
+            cross_attention_mask: Optional[torch.Tensor] = None,
+            position_ids: Optional[torch.LongTensor] = None,
+            past_key_values: Optional[List[torch.FloatTensor]] = None,
+            inputs_embeds: Optional[torch.FloatTensor] = None,
+            use_cache: Optional[bool] = None,
+            output_attentions: Optional[bool] = None,
+            output_hidden_states: Optional[bool] = None,
+            return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, BaseModelOutputWithPast]:
+        """take care of image_encode, token_type_ids, position_ids and (attention_mask = None is fine)"""
+
+        if past_key_values is not None:
+            encoder_outputs = None
+            # generate mode with past_key_values. the image features are already mapped
+        else:
+            # not allow for inputs_embeds, because we want to process image feature
+            assert input_ids is not None and inputs_embeds is None, f"{input_ids} {inputs_embeds}"
+            if not is_empty(images):  # multi-modality
+                assert token_type_ids is not None, f"multi-modality requires `token_type_ids`!"
+                assert len(input_ids) == len(images), f"{len(input_ids)} {len(images)}"
+                inputs_embeds = self.embed_tokens(input_ids)
+                images_features = self.encode_images(images)
+                encoder_outputs = self.encode_cross_images(cross_images)
+                images_features = rearrange(images_features, 'b n d -> (b n) d')
+                images_features = images_features.to(dtype=inputs_embeds.dtype, device=inputs_embeds.device)
+                inputs_embeds = inputs_embeds.index_put([token_type_ids == VISION_TOKEN_TYPE], images_features)
+            else:  # single-modality
+                if token_type_ids is None:
+                    token_type_ids = torch.ones_like(input_ids, dtype=torch.long, device=input_ids.device) * LANGUAGE_TOKEN_TYPE
+                assert not (token_type_ids == VISION_TOKEN_TYPE).any(), f"{(token_type_ids == VISION_TOKEN_TYPE).sum()}"
+                inputs_embeds = self.embed_tokens(input_ids)
+                encoder_outputs = None
+
+            if position_ids is None:
+                position_ids = build_position_ids(token_type_ids, attention_mask)
+            input_ids = None
+
+        return self.llm_forward(
+            input_ids=input_ids,
+            encoder_outputs=encoder_outputs,
+            token_type_ids=token_type_ids,
+            attention_mask=attention_mask,
+            cross_attention_mask=cross_attention_mask,
+            position_ids=position_ids,
+            past_key_values=past_key_values,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+    def llm_forward(
+            self,
+            input_ids: torch.LongTensor = None,
+            encoder_outputs: torch.LongTensor = None,
+            token_type_ids: torch.LongTensor = None,
+            attention_mask: Optional[torch.Tensor] = None,
+            cross_attention_mask: Optional[torch.Tensor] = None,
+            position_ids: Optional[torch.LongTensor] = None,
+            past_key_values: Optional[List[torch.FloatTensor]] = None,
+            inputs_embeds: Optional[torch.FloatTensor] = None,
+            use_cache: Optional[bool] = None,
+            output_attentions: Optional[bool] = None,
+            output_hidden_states: Optional[bool] = None,
+            return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, BaseModelOutputWithPast]:
+        """largely copy from llama forward and adapt for CogAgent with `token_type_ids`"""
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        use_cache = use_cache if use_cache is not None else self.config.use_cache
+
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        # retrieve input_ids and inputs_embeds
+        if input_ids is not None and inputs_embeds is not None:
+            raise ValueError("You cannot specify both decoder_input_ids and decoder_inputs_embeds at the same time")
+        elif input_ids is not None:
+            batch_size, seq_length = input_ids.shape
+        elif inputs_embeds is not None:
+            batch_size, seq_length, _ = inputs_embeds.shape
+        else:
+            raise ValueError("You have to specify either decoder_input_ids or decoder_inputs_embeds")
+
+        seq_length_with_past = seq_length
+        past_key_values_length = 0
+
+        if past_key_values is not None:
+            past_key_values_length = past_key_values[0][0].shape[2]
+            seq_length_with_past = seq_length_with_past + past_key_values_length
+
+        if position_ids is None:
+            device = input_ids.device if input_ids is not None else inputs_embeds.device
+            position_ids = torch.arange(
+                past_key_values_length, seq_length + past_key_values_length, dtype=torch.long, device=device
+            )
+            position_ids = position_ids.unsqueeze(0).view(-1, seq_length)
+        else:
+            position_ids = position_ids.view(-1, seq_length).long()
+
+        if inputs_embeds is None:
+            inputs_embeds = self.embed_tokens(input_ids)
+        # embed positions
+        if attention_mask is None:
+            attention_mask = torch.ones(
+                (batch_size, seq_length_with_past), dtype=torch.bool, device=inputs_embeds.device
+            )
+        if cross_attention_mask is None:
+            cross_attention_mask = torch.ones(
+                (batch_size, 1), dtype=torch.bool, device=inputs_embeds.device
+            )
+        attention_mask = self._prepare_decoder_attention_mask(
+            attention_mask, (batch_size, seq_length), inputs_embeds, past_key_values_length
+        )
+
+        hidden_states = inputs_embeds
+
+        # decoder layers
+        all_hidden_states = () if output_hidden_states else None
+        all_self_attns = () if output_attentions else None
+        next_decoder_cache = () if use_cache else None
+
+        for idx, decoder_layer in enumerate(self.layers):
+            if output_hidden_states:
+                all_hidden_states += (hidden_states,)
+
+            past_key_value = past_key_values[idx] if past_key_values is not None else None
+            layer_outputs = decoder_layer(
+                hidden_states,
+                encoder_outputs=encoder_outputs,
+                token_type_ids=token_type_ids,
+                attention_mask=attention_mask,
+                cross_attention_mask=cross_attention_mask,
+                position_ids=position_ids,
+                past_key_value=past_key_value,
+                output_attentions=output_attentions,
+                use_cache=use_cache,
+            )
+            hidden_states = layer_outputs[0]
+
+            if use_cache:
+                next_decoder_cache += (layer_outputs[2 if output_attentions else 1],)
+
+            if output_attentions:
+                all_self_attns += (layer_outputs[1],)
+
+        hidden_states = self.norm(hidden_states)
+
+        # add hidden states from the last decoder layer
+        if output_hidden_states:
+            all_hidden_states += (hidden_states,)
+
+        next_cache = next_decoder_cache if use_cache else None
+        if not return_dict:
+            return tuple(v for v in [hidden_states, next_cache, all_hidden_states, all_self_attns] if v is not None)
+        return BaseModelOutputWithPast(
+            last_hidden_state=hidden_states,
+            past_key_values=next_cache,
+            hidden_states=all_hidden_states,
+            attentions=all_self_attns,
+        )
+
+    def get_input_embeddings(self):
+        return self.embed_tokens
+
+    def set_input_embeddings(self, value):
+        self.embed_tokens = value
+
+    # noinspection PyMethodMayBeStatic
+    # Copied from transformers.models.bart.modeling_bart.BartDecoder._prepare_decoder_attention_mask
+    def _prepare_decoder_attention_mask(self, attention_mask, input_shape, inputs_embeds, past_key_values_length):
+        # create causal mask
+        # [bsz, seq_len] -> [bsz, 1, tgt_seq_len, src_seq_len]
+        combined_attention_mask = None
+        if input_shape[-1] > 1:
+            combined_attention_mask = _make_causal_mask(
+                input_shape,
+                inputs_embeds.dtype,
+                device=inputs_embeds.device,
+                past_key_values_length=past_key_values_length,
+            )
+
+        if attention_mask is not None:
+            # [bsz, seq_len] -> [bsz, 1, tgt_seq_len, src_seq_len]
+            expanded_attn_mask = _expand_mask(attention_mask, inputs_embeds.dtype, tgt_len=input_shape[-1]).to(
+                inputs_embeds.device
+            )
+            combined_attention_mask = (
+                expanded_attn_mask if combined_attention_mask is None else expanded_attn_mask + combined_attention_mask
+            )
+
+        return combined_attention_mask
+
+def vqa_history_to_prompt(history, query):
+    # Only support single round chat in vqa mode
+    prompt = "<EOI>Question: "
+    # for i, (old_query, response) in enumerate(history):
+    #     prompt += old_query + " Short answer: " + response + " Question: "
+    prompt += query + " Short answer:"
+    return prompt
+
+def chat_old_history_to_prompt(history, query):
+    prompt = "<EOI>Question: "
+    for i, (old_query, response) in enumerate(history):
+        prompt += old_query + " Answer: " + response + "\nQuestion: "
+    prompt += query + " Answer:"
+    return prompt
+
+def chat_history_to_prompt(history, query):
+    prompt = " [INST] "
+    for i, (old_query, response) in enumerate(history):
+        prompt += old_query + " [/INST] " + response + " [INST] "
+    prompt += query + " [/INST] "
+    return prompt
+
+
+def base_history_to_prompt(history, query):
+    prompt = query
+    return prompt
+
+
+_history_to_prompt = {
+    "base": base_history_to_prompt,
+    "chat": chat_history_to_prompt,
+    "chat_old": chat_old_history_to_prompt,
+    "vqa": vqa_history_to_prompt
+}
+
+
+class CogAgentForCausalLM(CogAgentPreTrainedModel):
+    _auto_class = "AutoModelForCausalLM"
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.model = CogAgentModel(config)
+        self.vocab_size = config.vocab_size
+        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_input_embeddings(self):
+        return self.model.embed_tokens
+
+    def set_input_embeddings(self, value):
+        self.model.embed_tokens = value
+
+    def get_output_embeddings(self):
+        return self.lm_head
+
+    def set_output_embeddings(self, new_embeddings):
+        self.lm_head = new_embeddings
+
+    def set_decoder(self, decoder):
+        self.model = decoder
+
+    def get_decoder(self):
+        return self.model
+
+    def forward(
+            self,
+            input_ids: torch.LongTensor = None,
+            images: List[List[torch.Tensor]] = None,
+            cross_images: List[List[torch.Tensor]] = None,
+            token_type_ids: Optional[torch.LongTensor] = None,
+            attention_mask: Optional[torch.Tensor] = None,
+            position_ids: Optional[torch.LongTensor] = None,
+            past_key_values: Optional[List[torch.FloatTensor]] = None,
+            inputs_embeds: Optional[torch.FloatTensor] = None,
+            use_cache: Optional[bool] = None,
+            output_attentions: Optional[bool] = None,
+            output_hidden_states: Optional[bool] = None,
+            return_dict: Optional[bool] = None,
+            labels: Optional[torch.LongTensor] = None,
+    ) -> Union[Tuple, CausalLMOutputWithPast]:
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        # decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)
+        outputs = self.model(
+            input_ids=input_ids,
+            images=images,
+            cross_images=cross_images,
+            token_type_ids=token_type_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            past_key_values=past_key_values,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        hidden_states = outputs[0]
+        logits = self.lm_head(hidden_states)
+        logits = logits.float()
+
+        loss = None
+        if labels is not None:
+            # Shift so that tokens < n predict n
+            shift_logits = logits[..., :-1, :].contiguous()
+            shift_labels = labels[..., 1:].contiguous()
+            # Flatten the tokens
+            loss_fct = CrossEntropyLoss()
+            shift_logits = shift_logits.view(-1, self.config.vocab_size)
+            shift_labels = shift_labels.view(-1)
+            # Enable model parallelism
+            shift_labels = shift_labels.to(shift_logits.device)
+            loss = loss_fct(shift_logits, shift_labels)
+
+        if not return_dict:
+            output = (logits,) + outputs[1:]
+            return (loss,) + output if loss is not None else output
+
+        return CausalLMOutputWithPast(
+            loss=loss,
+            logits=logits,
+            past_key_values=outputs.past_key_values,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+    def _prepare_attention_mask_for_generation(
+            self,
+            inputs: torch.Tensor,
+            pad_token_id: Optional[int],
+            eos_token_id: Optional[Union[int, List[int]]],
+    ) -> torch.LongTensor:
+        return torch.ones(inputs.shape[:2], dtype=torch.long, device=inputs.device)  # type: ignore
+
+    def prepare_inputs_for_generation(
+            self, input_ids, token_type_ids, images=None, cross_images=None, past_key_values=None, attention_mask=None, inputs_embeds=None, **kwargs
+    ):
+        # build position_ids if needed
+        position_ids = kwargs.get("position_ids", None)
+        if position_ids is None:
+            position_ids = build_position_ids(token_type_ids, attention_mask)
+
+        if past_key_values:
+            input_ids = input_ids[:, -1:]
+            token_type_ids = token_type_ids[:, -1:]
+            position_ids = position_ids[:, -1:]
+
+        # if `inputs_embeds` are passed, we only want to use them in the 1st generation step
+        if inputs_embeds is not None and past_key_values is None:
+            model_inputs = {"inputs_embeds": inputs_embeds}
+        else:
+            model_inputs = {"input_ids": input_ids}
+
+        model_inputs.update(
+            {
+                "token_type_ids": token_type_ids,
+                "images": images,
+                "cross_images": cross_images,
+                "position_ids": position_ids,
+                "past_key_values": past_key_values,
+                "use_cache": kwargs.get("use_cache"),
+                "attention_mask": attention_mask,
+            }
+        )
+        return model_inputs
+
+    def _update_model_kwargs_for_generation(
+            self,
+            outputs: "ModelOutput",
+            model_kwargs: Dict[str, Any],
+            is_encoder_decoder: bool = False,
+            standardize_cache_format: bool = False,
+    ) -> Dict[str, Any]:
+        # update past_key_values
+        model_kwargs["past_key_values"] = self._extract_past_from_model_output(
+            outputs, standardize_cache_format=standardize_cache_format
+        )
+        if getattr(outputs, "state", None) is not None:
+            model_kwargs["state"] = outputs.state
+
+        # update token_type_ids with last value
+        if "token_type_ids" in model_kwargs:
+            token_type_ids = model_kwargs["token_type_ids"]
+            new_token_type_ids = torch.ones(size=(token_type_ids.shape[0], 1), dtype=token_type_ids.dtype, device=token_type_ids.device) * LANGUAGE_TOKEN_TYPE
+            model_kwargs["token_type_ids"] = torch.cat([token_type_ids, new_token_type_ids], dim=-1)
+
+        if not is_encoder_decoder:
+            # update attention mask
+            if "attention_mask" in model_kwargs:
+                attention_mask = model_kwargs["attention_mask"]
+                model_kwargs["attention_mask"] = torch.cat(
+                    [attention_mask, attention_mask.new_ones((attention_mask.shape[0], 1))], dim=-1
+                )
+        else:
+            # update decoder attention mask
+            if "decoder_attention_mask" in model_kwargs:
+                decoder_attention_mask = model_kwargs["decoder_attention_mask"]
+                model_kwargs["decoder_attention_mask"] = torch.cat(
+                    [decoder_attention_mask, decoder_attention_mask.new_ones((decoder_attention_mask.shape[0], 1))],
+                    dim=-1,
+                )
+
+        return model_kwargs
+
+    def _reorder_cache(self, past_key_values, beam_idx):
+        reordered_past = ()
+        for layer_past in past_key_values:
+            reordered_past += (
+                tuple(past_state.index_select(0, beam_idx.to(past_state.device)) for past_state in layer_past),
+            )
+        return reordered_past
+
+    def build_conversation_input_ids(
+            self,
+            tokenizer: "PreTrainedTokenizer",
+            *,
+            query: str,
+            history: Optional[List[Tuple[str, str]]] = None,
+            images: Optional[List["PIL.Image"]] = None,
+            template_version: Optional[Literal["base", "chat", "vqa"]] = None,
+    ):
+        image_size: int = self.config.vision_config['image_size']
+        cross_image_size: int = self.config.cross_image_size
+        patch_size: int = self.config.vision_config['patch_size']
+        template_version = template_version or self.config.template_version
+        assert images is None or len(images) <= 1, f"not support multi images by now."
+        history = history or []
+        text = _history_to_prompt[template_version](history, query)
+
+        input_ids = [tokenizer.bos_token_id]
+        token_type_ids = [LANGUAGE_TOKEN_TYPE]
+        if images is not None and len(images) == 1:
+            ori = images
+            # vision
+            transform = transforms.Compose(
+                [
+                    transforms.Resize(
+                        (image_size, image_size), interpolation=transforms.InterpolationMode.BICUBIC
+                    ),
+                    transforms.ToTensor(),
+                    transforms.Normalize((0.48145466, 0.4578275, 0.40821073), (0.26862954, 0.26130258, 0.27577711)),
+                ]
+            )
+            images = [transform(ori[0])]
+            cross_transform = transforms.Compose(
+                [
+                    transforms.Resize(
+                        (cross_image_size, cross_image_size), interpolation=transforms.InterpolationMode.BICUBIC
+                    ),
+                    transforms.ToTensor(),
+                    transforms.Normalize((0.48145466, 0.4578275, 0.40821073), (0.26862954, 0.26130258, 0.27577711)),
+                ]
+            )
+            cross_images = [cross_transform(ori[0])]
+            # language
+            vision_token_num = (image_size // patch_size) * (image_size // patch_size) + 2
+            input_ids += [tokenizer.pad_token_id] * vision_token_num
+            token_type_ids += [VISION_TOKEN_TYPE] * vision_token_num
+        text_ids = tokenizer.encode(text, add_special_tokens=False)
+
+        input_ids += text_ids
+        token_type_ids += [LANGUAGE_TOKEN_TYPE] * len(text_ids)
+        attention_mask = [1] * len(input_ids)
+
+        return {
+            'input_ids': torch.tensor(input_ids, dtype=torch.long),
+            'token_type_ids': torch.tensor(token_type_ids, dtype=torch.long),
+            'attention_mask': torch.tensor(attention_mask, dtype=torch.long),
+            'images': images,
+            'cross_images': cross_images
+        }