Merge branch 'main' into moonride-main

Author: MoonRide303

colab.ipynb

@@ -10,29 +10,10 @@
       "source": [
         "%cd /content\n",
         "!git clone https://github.com/MoonRide303/Fooocus-MRE.git\n",
-        "!cp /content/Fooocus-MRE/settings-no-refiner.json /content/Fooocus-MRE/settings.json\n",
-        "\n",
-        "!apt -y update -qq\n",
-        "!wget https://github.com/camenduru/gperftools/releases/download/v1.0/libtcmalloc_minimal.so.4 -O /content/libtcmalloc_minimal.so.4\n",
-        "%env LD_PRELOAD=/content/libtcmalloc_minimal.so.4\n",
-        "\n",
-        "!pip install torchsde==0.2.5 einops==0.4.1 transformers==4.30.2 safetensors==0.3.1 accelerate==0.21.0 aiohttp==3.8.5\n",
-        "!pip install pytorch_lightning==1.9.4 omegaconf==2.2.3 gradio==3.41.2 xformers==0.0.21 triton==2.0.0 pygit2==1.12.2 fastapi==0.94.0\n",
-        "\n",
-        "!apt -y install -qq aria2\n",
-        "!aria2c --console-log-level=error -c -x 16 -s 16 -k 1M https://huggingface.co/ckpt/sd_xl_base_1.0/resolve/main/sd_xl_base_1.0_0.9vae.safetensors -d /content/Fooocus-MRE/models/checkpoints -o sd_xl_base_1.0_0.9vae.safetensors\n",
-        "!aria2c --console-log-level=error -c -x 16 -s 16 -k 1M https://huggingface.co/ckpt/sd_xl_refiner_1.0/resolve/main/sd_xl_refiner_1.0_0.9vae.safetensors -d /content/Fooocus-MRE/models/checkpoints -o sd_xl_refiner_1.0_0.9vae.safetensors\n",
-        "!aria2c --console-log-level=error -c -x 16 -s 16 -k 1M https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/resolve/main/sd_xl_offset_example-lora_1.0.safetensors -d /content/Fooocus-MRE/models/loras -o sd_xl_offset_example-lora_1.0.safetensors\n",
-        "!aria2c --console-log-level=error -c -x 16 -s 16 -k 1M https://huggingface.co/stabilityai/control-lora/resolve/main/revision/clip_vision_g.safetensors -d /content/Fooocus-MRE/models/clip_vision -o clip_vision_g.safetensors\n",
-        "!aria2c --console-log-level=error -c -x 16 -s 16 -k 1M https://huggingface.co/stabilityai/control-lora/resolve/main/control-LoRAs-rank128/control-lora-canny-rank128.safetensors -d /content/Fooocus-MRE/models/controlnet -o control-lora-canny-rank128.safetensors\n",
-        "!aria2c --console-log-level=error -c -x 16 -s 16 -k 1M https://huggingface.co/stabilityai/control-lora/resolve/main/control-LoRAs-rank256/control-lora-canny-rank256.safetensors -d /content/Fooocus-MRE/models/controlnet -o control-lora-canny-rank256.safetensors\n",
-        "!aria2c --console-log-level=error -c -x 16 -s 16 -k 1M https://huggingface.co/stabilityai/control-lora/resolve/main/control-LoRAs-rank128/control-lora-depth-rank128.safetensors -d /content/Fooocus-MRE/models/controlnet -o control-lora-depth-rank128.safetensors\n",
-        "!aria2c --console-log-level=error -c -x 16 -s 16 -k 1M https://huggingface.co/stabilityai/control-lora/resolve/main/control-LoRAs-rank256/control-lora-depth-rank256.safetensors -d /content/Fooocus-MRE/models/controlnet -o control-lora-depth-rank256.safetensors\n",
-
-        "\n",
-        "%cd /content/Fooocus-MRE\n",
-        "!git pull\n",
-        "!python launch.py --share\n"
+        "%cd Fooocus-MRE\n",
+        "!cp settings-no-refiner.json settings.json\n",
+        "!pip install pygit2==1.12.2\n",
+        "!python entry_with_update.py --share\n"
       ]
     }
   ],

fooocus_version.py

@@ -1,2 +1,2 @@
-version = '2.0.8 MRE'
+version = '2.0.12 MRE'
 full_version = 'Fooocus ' + version

models/prompt_expansion/fooocus_expansion/config.json

@@ -8,6 +8,7 @@
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
+  "pad_token_id": 50256,
   "initializer_range": 0.02,
   "layer_norm_epsilon": 1e-05,
   "model_type": "gpt2",

modules/expansion.py

@@ -1,8 +1,7 @@
-import torch
 import comfy.model_management as model_management
 import modules.constants as constants
 
-from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline, set_seed
+from transformers import AutoTokenizer, AutoModelForCausalLM, set_seed
 from modules.path import fooocus_expansion_path
 from comfy.model_patcher import ModelPatcher
 
@@ -29,35 +28,38 @@ def remove_pattern(x, pattern):
 
 class FooocusExpansion:
     def __init__(self):
-        use_fp16 = model_management.should_use_fp16()
-
         self.tokenizer = AutoTokenizer.from_pretrained(fooocus_expansion_path)
         self.model = AutoModelForCausalLM.from_pretrained(fooocus_expansion_path)
 
-        if use_fp16:
+        if model_management.should_use_fp16():
             self.model.half()
 
         load_device = model_management.text_encoder_device()
         offload_device = model_management.text_encoder_offload_device()
         self.patcher = ModelPatcher(self.model, load_device=load_device, offload_device=offload_device)
 
-        self.pipe = pipeline('text-generation',
-                             model=self.model,
-                             tokenizer=self.tokenizer,
-                             device='cpu',
-                             torch_dtype=torch.float16 if use_fp16 else torch.float32)
-
         print(f'Fooocus Expansion engine loaded.')
 
     def __call__(self, prompt, seed):
         model_management.load_model_gpu(self.patcher)
-        self.pipe.device = self.patcher.load_device
         seed = int(seed) % constants.SEED_LIMIT_NUMPY
         set_seed(seed)
         origin = safe_str(prompt)
         prompt = origin + fooocus_magic_split[seed % len(fooocus_magic_split)]
-        response = self.pipe(prompt, max_length=len(prompt) + 256)
-        result = response[0]['generated_text'][len(origin):]
+
+        tokenized_kwargs = self.tokenizer(prompt, return_tensors="pt")
+        tokenized_kwargs.data['input_ids'] = tokenized_kwargs.data['input_ids'].to(self.patcher.load_device)
+        tokenized_kwargs.data['attention_mask'] = tokenized_kwargs.data['attention_mask'].to(self.patcher.load_device)
+
+        # https://huggingface.co/blog/introducing-csearch
+        # https://huggingface.co/docs/transformers/generation_strategies
+        features = self.model.generate(**tokenized_kwargs,
+                                       num_beams=1,
+                                       max_new_tokens=256,
+                                       do_sample=True)
+
+        response = self.tokenizer.batch_decode(features, skip_special_tokens=True)
+        result = response[0][len(origin):]
         result = safe_str(result)
         result = remove_pattern(result, dangrous_patterns)
         return result

readme.md

@@ -47,6 +47,8 @@ Below is a test on a relatively low-end laptop with **16GB System RAM** and **6G
 
 ![image](https://github.com/lllyasviel/Fooocus/assets/19834515/938737a5-b105-4f19-b051-81356cb7c495)
 
+Besides, recently many other software report that Nvidia driver above 532 is sometimes 10x slower than Nvidia driver 531. If your generation time is very long, consider download [Nvidia Driver 531 Laptop](https://www.nvidia.com/download/driverResults.aspx/199991/en-us/) or [Nvidia Driver 531 Desktop](https://www.nvidia.com/download/driverResults.aspx/199990/en-us/).
+
 Note that the minimal requirement is **4GB Nvidia GPU memory (4GB VRAM)** and **8GB system memory (8GB RAM)**. This requires using Microsoft’s Virtual Swap technique, which is automatically enabled by your Windows installation in most cases, so you often do not need to do anything about it. However, if you are not sure, or if you manually turned it off (would anyone really do that?), or **if you see any "RuntimeError: CPUAllocator"**, you can enable it here:
 
 <details>
@@ -69,9 +71,7 @@ Please open an issue if you use similar devices but still cannot achieve accepta
 [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/lllyasviel/Fooocus/blob/main/colab.ipynb) | Fooocus Colab (Official Version)
 [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/MoonRide303/Fooocus-MRE/blob/moonride-main/colab.ipynb) | Fooocus-MRE Colab (MoonRide Edition)
 
-Note that sometimes this Colab will say like "you must restart the runtime in order to use newly installed XX". This can be safely ignored.
-
-Thanks to [camenduru](https://github.com/camenduru)'s codes!
+Thanks to [camenduru](https://github.com/camenduru)!
 
 ### Linux