nodes/MinusZoneAI-ComfyUI-Prompt-MZ
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

'update'

Browse Source
This commit is contained in:
wailovet
2024-04-30 12:46:04 +08:00
parent f0122b52bf d5c64e056a
commit d5ad49de32
4 changed files with 27 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+19 -9
View File
@@ -1,34 +1,44 @@
![Image Description](res/1.png)
# ComfyUI-Prompt-MZ
基于llama.cpp的一些和提示词相关的节点，目前包括美化提示词和类似clip-interrogator的图片反推
Use llama.cpp to assist in generating some nodes related to prompt words, including beautifying prompt words and image recognition similar to clip-interrogator
## Recent changes
* [2024-04-30] 添加llava-v1.6-vicuna-13b (add llava-v1.6-vicuna-13b)
* [2024-04-30] 添加翻译
* [2024-04-28] 新增Phi-3-mini节点 (add Phi-3-mini node)
## Installation
1. Clone this repo into `custom_nodes` folder.
2. Restart ComfyUI.
## Nodes
+ LLama3CLIPTextEncode
+ CLIPTextEncode (Phi-3)
![image](https://github.com/MinusZoneAI/ComfyUI-Prompt-MZ/assets/5035199/c4b97aeb-23c0-4cf1-a6a5-d259fdf83f6e)
+ CLIPTextEncode (LLama3)
![image](https://github.com/MinusZoneAI/ComfyUI-Prompt-MZ/assets/5035199/40da75ab-46db-4f38-9d8e-b7f9184f77fa)
+ LLavaImageInterrogator
+ ImageInterrogator (LLava)
![image](https://github.com/MinusZoneAI/ComfyUI-Prompt-MZ/assets/5035199/f397c432-c2f7-4d48-9b95-2031cfb19e8c)
Enable parameter sd_format
![image](https://github.com/MinusZoneAI/ComfyUI-Prompt-MZ/assets/5035199/4d2cf65d-e8a3-4dfa-b735-9d591638028c)
+ LLamaCPPOptions
![image](https://github.com/MinusZoneAI/ComfyUI-Prompt-MZ/assets/5035199/143d2055-31fa-4130-9937-1c252bda9c73)
![image](https://github.com/MinusZoneAI/ComfyUI-Prompt-MZ/assets/5035199/256483e0-c3b7-4d04-82f4-f71f7d9584c9)
+ BaseLLamaCPPCLIPTextEncode (You can directly pass in the model path)
+ BaseLLavaImageInterrogator (You can directly pass in the model path)
+ BaseLLamaCPPCLIPTextEncode (可以手动传入模型路径/You can directly pass in the model path)
+ BaseLLavaImageInterrogator (可以手动传入模型路径/You can directly pass in the model path)
mz_llama_cpp.py
+1 -1
View File
@@ -351,7 +351,7 @@ def llava_cpp_messages(model_file, chat_handler, messages, options={}):
if options is None:
options = {}
options["logits_all"] = True
options["n_ctx"] = max(2048, options.get("n_ctx", 2048))
options["n_ctx"] = max(4096, options.get("n_ctx", 4096))
return llama_cpp_messages(model_file, chat_handler, messages, options)
mz_llava.py
+6 -4
View File
@@ -9,9 +9,9 @@ import mz_prompts
LLava_models = [
"llava-1.6-mistral-7b-gguf/llava-1.6-mistral-7b.Q5_K_M.gguf",
"llava-1.6-mistral-7b-gguf/llava-v1.6-mistral-7b.Q5_K_M.gguf",
"llava-v1.6-vicuna-13b-gguf/llava-v1.6-vicuna-13b.Q5_K_M.gguf",
"ggml_llava-v1.5-7b/ggml-model-q4_k.gguf",
"ggml_llava-v1.5-7b/ggml-model-q4_k.gguf",
"ggml_llava-v1.5-7b/ggml-model-q5_k.gguf",
"ggml_llava-v1.5-7b/ggml-model-f16.gguf",
"ggml_bakllava-1/ggml-model-q4_k.gguf",
@@ -154,17 +154,19 @@ def base_image_interrogator(args_dict):
schema = mz_llama_cpp.get_schema_obj(
keys_type={
"short_desc": mz_llama_cpp.get_schema_base_type("string"),
"short_describes": mz_llama_cpp.get_schema_base_type("string"),
"subject_tags": mz_llama_cpp.get_schema_array("string"),
"action_tags": mz_llama_cpp.get_schema_array("string"),
"light_tags": mz_llama_cpp.get_schema_array("string"),
"scenes_tags": mz_llama_cpp.get_schema_array("string"),
"other_tags": mz_llama_cpp.get_schema_array("string"),
},
required=[
"short_desc",
"short_describes",
"subject_tags",
"action_tags",
"light_tags",
"scenes_tags",
"other_tags",
]
)
mz_prompt_utils.py
+1 -1
View File
@@ -588,7 +588,7 @@ modelscope_models_map = {
"output": "modelscope_models/llava-v1.6-vicuna-13b-gguf/mmproj-model-f16.gguf",
"SHA256": "9860fbc162d6df40d12f32211aed11ab4826e16e473e6fff595496b9943872e2"
},
"llava-1.6-mistral-7b-gguf/llava-1.6-mistral-7b.Q5_K_M.gguf": {
"llava-1.6-mistral-7b-gguf/llava-v1.6-mistral-7b.Q5_K_M.gguf": {
"url": "https://www.modelscope.cn/api/v1/models/wailovet/MinusZoneAIModels/repo?Revision=master&FilePath=llava-1.6-mistral-7b-gguf%2Fllava-v1.6-mistral-7b.Q5_K_M.gguf",
"output": "modelscope_models/llava-1.6-mistral-7b-gguf/llava-v1.6-mistral-7b.Q5_K_M.gguf",
"SHA256": "b1d37fc65ecb80aa8f1ce185bf4d7605bc3c5cc5bcc77a160c3a1b0377631112",