torch.compile won't work yet, keep code prepared

4 files changed, 52 insertions, 4 deletions

diff --git a/environment.yaml b/environment.yaml
index b161244..dfbafaf 100644
--- a/environment.yaml
+++ b/environment.yaml
@@ -4,7 +4,7 @@ channels:
     - nvidia
     - xformers/label/dev
     - defaults
-    - conda-forge 
+    - conda-forge
 dependencies:
     - gcc=11.3.0
     - gxx=11.3.0
@@ -30,4 +30,3 @@ dependencies:
           - test-tube>=0.7.5
           - timm==0.8.17.dev0
           - transformers==4.28.1
-          - triton==2.0.0.post1
diff --git a/environment_nightly.yaml b/environment_nightly.yaml
new file mode 100644
index 0000000..4c5c798
--- /dev/null
+++ b/environment_nightly.yaml
@@ -0,0 +1,31 @@
+name: ldd
+channels:
+    - pytorch-nightly
+    - nvidia
+    - xformers/label/dev
+    - defaults
+    - conda-forge 
+dependencies:
+    - cuda-nvcc=12.1.105
+    - matplotlib=3.6.2
+    - numpy=1.24.3
+    - pip=22.3.1
+    - python=3.10.8
+    - pytorch=2.1.0.dev20230429=*cuda12.1*
+    - torchvision=0.16.0.dev20230429
+    # - xformers=0.0.19
+    - pip:
+          - -e .
+          - -e git+https://github.com/huggingface/accelerate#egg=accelerate
+          - -e git+https://github.com/huggingface/diffusers#egg=diffusers
+          - -e git+https://github.com/facebookresearch/dadaptation#egg=dadaptation
+          - bitsandbytes==0.38.1
+          - hidet==0.2.3
+          - lion-pytorch==0.0.7
+          - peft==0.2.0
+          - python-slugify>=6.1.2
+          - safetensors==0.3.1
+          - setuptools==65.6.3
+          - test-tube>=0.7.5
+          - timm==0.8.17.dev0
+          - transformers==4.28.1
diff --git a/train_lora.py b/train_lora.py
index d95dbb9..3c8fc97 100644
--- a/train_lora.py
+++ b/train_lora.py
@@ -16,6 +16,7 @@ from accelerate import Accelerator
 from accelerate.logging import get_logger
 from accelerate.utils import LoggerType, set_seed
 from peft import LoraConfig, LoraModel
+# from diffusers.models.attention_processor import AttnProcessor
 import transformers
 
 import numpy as np
@@ -41,10 +42,11 @@ warnings.filterwarnings('ignore')
 torch.backends.cuda.matmul.allow_tf32 = True
 torch.backends.cudnn.benchmark = True
 
-torch._dynamo.config.log_level = logging.ERROR
+torch._dynamo.config.log_level = logging.WARNING
 
 hidet.torch.dynamo_config.use_tensor_core(True)
-hidet.torch.dynamo_config.search_space(1)
+# hidet.torch.dynamo_config.use_attention(True)
+hidet.torch.dynamo_config.search_space(0)
 
 
 def parse_args():
@@ -724,6 +726,19 @@ def main():
     if args.use_xformers:
         vae.set_use_memory_efficient_attention_xformers(True)
         unet.enable_xformers_memory_efficient_attention()
+    # elif args.compile_unet:
+    #     unet.mid_block.attentions[0].transformer_blocks[0].attn1._use_2_0_attn = False
+    #
+    #     proc = AttnProcessor()
+    #
+    #     def fn_recursive_set_proc(module: torch.nn.Module):
+    #         if hasattr(module, "processor"):
+    #             module.processor = proc
+    #
+    #         for child in module.children():
+    #             fn_recursive_set_proc(child)
+    #
+    #     fn_recursive_set_proc(unet)
 
     if args.gradient_checkpointing:
         unet.enable_gradient_checkpointing()
diff --git a/training/functional.py b/training/functional.py
index 68ea40c..38dd59f 100644
--- a/training/functional.py
+++ b/training/functional.py
@@ -700,8 +700,11 @@ def train(
     vae.requires_grad_(False)
     vae.eval()
 
+    vae = torch.compile(vae, backend='hidet')
+
     if compile_unet:
         unet = torch.compile(unet, backend='hidet')
+        # unet = torch.compile(unet)
 
     callbacks = strategy.callbacks(
         accelerator=accelerator,