Allow downloading single file from UI (#3737)

README.md

@@ -164,8 +164,6 @@ text-generation-webui
 │   ├── llama-13b.ggmlv3.q4_K_M.bin
 ```
 
-Those models must be downloaded manually, as they are not currently supported by the automated downloader.
-
 #### GPT-4chan
 
 <details>

download-model.py

@@ -73,7 +73,7 @@ class ModelDownloader:
 
             for i in range(len(dict)):
                 fname = dict[i]['path']
-                if specific_file is not None and fname != specific_file:
+                if specific_file not in [None, ''] and fname != specific_file:
                     continue
 
                 if not is_lora and fname.endswith(('adapter_config.json', 'adapter_model.bin')):
@@ -175,26 +175,28 @@ class ModelDownloader:
                         f.write(data)
                         if total_size != 0 and self.progress_bar is not None:
                             count += len(data)
-                            self.progress_bar(float(count) / float(total_size), f"Downloading {filename}")
+                            self.progress_bar(float(count) / float(total_size), f"{filename}")
 
     def start_download_threads(self, file_list, output_folder, start_from_scratch=False, threads=1):
         thread_map(lambda url: self.get_single_file(url, output_folder, start_from_scratch=start_from_scratch), file_list, max_workers=threads, disable=True)
 
-    def download_model_files(self, model, branch, links, sha256, output_folder, progress_bar=None, start_from_scratch=False, threads=1, specific_file=None):
+    def download_model_files(self, model, branch, links, sha256, output_folder, progress_bar=None, start_from_scratch=False, threads=1, specific_file=None, is_llamacpp=False):
         self.progress_bar = progress_bar
 
-        # Creating the folder and writing the metadata
+        # Create the folder and writing the metadata
         output_folder.mkdir(parents=True, exist_ok=True)
-        metadata = f'url: https://huggingface.co/{model}\n' \
-                   f'branch: {branch}\n' \
-                   f'download date: {datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")}\n'
 
-        sha256_str = '\n'.join([f'    {item[1]} {item[0]}' for item in sha256])
+        if not is_llamacpp:
-        if sha256_str:
+            metadata = f'url: https://huggingface.co/{model}\n' \
-            metadata += f'sha256sum:\n{sha256_str}'
+                       f'branch: {branch}\n' \
+                       f'download date: {datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")}\n'
 
-        metadata += '\n'
+            sha256_str = '\n'.join([f'    {item[1]} {item[0]}' for item in sha256])
-        (output_folder / 'huggingface-metadata.txt').write_text(metadata)
+            if sha256_str:
+                metadata += f'sha256sum:\n{sha256_str}'
+
+            metadata += '\n'
+            (output_folder / 'huggingface-metadata.txt').write_text(metadata)
 
         if specific_file:
             print(f"Downloading {specific_file} to {output_folder}")
@@ -270,4 +272,4 @@ if __name__ == '__main__':
         downloader.check_model_files(model, branch, links, sha256, output_folder)
     else:
         # Download files
-        downloader.download_model_files(model, branch, links, sha256, output_folder, specific_file=specific_file, threads=args.threads)
+        downloader.download_model_files(model, branch, links, sha256, output_folder, specific_file=specific_file, threads=args.threads, is_llamacpp=is_llamacpp)

modules/ui_model_menu.py

@@ -3,6 +3,7 @@ import math
 import re
 import traceback
 from functools import partial
+from pathlib import Path
 
 import gradio as gr
 import psutil
@@ -119,14 +120,17 @@ def create_ui():
                             shared.gradio['gptq_for_llama_info'] = gr.Markdown('GPTQ-for-LLaMa support is currently only kept for compatibility with older GPUs. AutoGPTQ or ExLlama is preferred when compatible. GPTQ-for-LLaMa is installed by default with the webui on supported systems. Otherwise, it has to be installed manually following the instructions here: [instructions](https://github.com/oobabooga/text-generation-webui/blob/main/docs/GPTQ-models-(4-bit-mode).md#installation-1).')
                             shared.gradio['exllama_info'] = gr.Markdown('For more information, consult the [docs](https://github.com/oobabooga/text-generation-webui/blob/main/docs/ExLlama.md).')
                             shared.gradio['exllama_HF_info'] = gr.Markdown('ExLlama_HF is a wrapper that lets you use ExLlama like a Transformers model, which means it can use the Transformers samplers. It\'s a bit slower than the regular ExLlama.')
-                            shared.gradio['llamacpp_HF_info'] = gr.Markdown('llamacpp_HF is a wrapper that lets you use llama.cpp like a Transformers model, which means it can use the Transformers samplers. To use it, make sure to first download oobabooga/llama-tokenizer under "Download custom model or LoRA".')
+                            shared.gradio['llamacpp_HF_info'] = gr.Markdown('llamacpp_HF is a wrapper that lets you use llama.cpp like a Transformers model, which means it can use the Transformers samplers. To use it, make sure to first download oobabooga/llama-tokenizer under "Download model or LoRA".')
 
             with gr.Column():
                 with gr.Row():
                     shared.gradio['autoload_model'] = gr.Checkbox(value=shared.settings['autoload_model'], label='Autoload the model', info='Whether to load the model as soon as it is selected in the Model dropdown.')
 
-                shared.gradio['custom_model_menu'] = gr.Textbox(label="Download custom model or LoRA", info="Enter the Hugging Face username/model path, for instance: facebook/galactica-125m. To specify a branch, add it at the end after a \":\" character like this: facebook/galactica-125m:main")
+                shared.gradio['custom_model_menu'] = gr.Textbox(label="Download model or LoRA", info="Enter the Hugging Face username/model path, for instance: facebook/galactica-125m. To specify a branch, add it at the end after a \":\" character like this: facebook/galactica-125m:main. To download a single file, enter its name in the second box.")
-                shared.gradio['download_model_button'] = gr.Button("Download")
+                shared.gradio['download_specific_file'] = gr.Textbox(placeholder="File name (for GGUF/GGML)", show_label=False, max_lines=1)
+                with gr.Row():
+                    shared.gradio['download_model_button'] = gr.Button("Download", variant='primary')
+                    shared.gradio['get_file_list'] = gr.Button("Get file list")
 
                 with gr.Row():
                     shared.gradio['model_status'] = gr.Markdown('No model is loaded' if shared.model_name == 'None' else 'Ready')
@@ -170,7 +174,8 @@ def create_event_handlers():
         save_model_settings, gradio('model_menu', 'interface_state'), gradio('model_status'), show_progress=False)
 
     shared.gradio['lora_menu_apply'].click(load_lora_wrapper, gradio('lora_menu'), gradio('model_status'), show_progress=False)
-    shared.gradio['download_model_button'].click(download_model_wrapper, gradio('custom_model_menu'), gradio('model_status'), show_progress=True)
+    shared.gradio['download_model_button'].click(download_model_wrapper, gradio('custom_model_menu', 'download_specific_file'), gradio('model_status'), show_progress=True)
+    shared.gradio['get_file_list'].click(partial(download_model_wrapper, return_links=True), gradio('custom_model_menu', 'download_specific_file'), gradio('model_status'), show_progress=True)
     shared.gradio['autoload_model'].change(lambda x: gr.update(visible=not x), gradio('autoload_model'), gradio('load_model'))
 
 
@@ -206,7 +211,7 @@ def load_lora_wrapper(selected_loras):
     yield ("Successfuly applied the LoRAs")
 
 
-def download_model_wrapper(repo_id, progress=gr.Progress()):
+def download_model_wrapper(repo_id, specific_file, progress=gr.Progress(), return_links=False):
     try:
         downloader_module = importlib.import_module("download-model")
         downloader = downloader_module.ModelDownloader()
@@ -220,11 +225,15 @@ def download_model_wrapper(repo_id, progress=gr.Progress()):
         model, branch = downloader.sanitize_model_and_branch_names(model, branch)
 
         yield ("Getting the download links from Hugging Face")
-        links, sha256, is_lora = downloader.get_download_links_from_huggingface(model, branch, text_only=False)
+        links, sha256, is_lora, is_llamacpp = downloader.get_download_links_from_huggingface(model, branch, text_only=False, specific_file=specific_file)
+
+        if return_links:
+            yield '\n\n'.join([f"`{Path(link).name}`" for link in links])
+            return
 
         yield ("Getting the output folder")
         base_folder = shared.args.lora_dir if is_lora else shared.args.model_dir
-        output_folder = downloader.get_output_folder(model, branch, is_lora, base_folder=base_folder)
+        output_folder = downloader.get_output_folder(model, branch, is_lora, is_llamacpp=is_llamacpp, base_folder=base_folder)
 
         if check:
             progress(0.5)
@@ -232,8 +241,8 @@ def download_model_wrapper(repo_id, progress=gr.Progress()):
             downloader.check_model_files(model, branch, links, sha256, output_folder)
             progress(1.0)
         else:
-            yield (f"Downloading files to {output_folder}")
+            yield (f"Downloading file{'s' if len(links) > 1 else ''} to `{output_folder}/`")
-            downloader.download_model_files(model, branch, links, sha256, output_folder, progress_bar=progress, threads=1)
+            downloader.download_model_files(model, branch, links, sha256, output_folder, progress_bar=progress, threads=1, is_llamacpp=is_llamacpp)
             yield ("Done!")
     except:
         progress(1.0)