Upload 3 files

app.py

@@ -62,152 +62,119 @@ In particular, conversion of FLUX.1 or SD 3.5 is almost impossible in CPU space.
             with gr.Group():
                 sdxl_presets = gr.Radio(label="Presets", choices=list(sdxl_preset_dict.keys()), value=list(sdxl_preset_dict.keys())[0])
                 sdxl_mtype = gr.Textbox(value="SDXL", visible=False)
-                with gr.Row():
-                    sdxl_dtype = gr.Radio(label="Output data type", choices=get_dtypes(), value=DEFAULT_DTYPE)
+                sdxl_dtype = gr.Radio(label="Output data type", choices=get_dtypes(), value=DEFAULT_DTYPE)
                 with gr.Accordion("Advanced settings", open=False):
                     with gr.Row():
                         sdxl_vae = gr.Dropdown(label="VAE", choices=sdxl_vaes, value="", allow_custom_value=True)
-                        sdxl_clip = gr.Dropdown(label="CLIP", choices=clips, value="", allow_custom_value=True)
                         sdxl_scheduler = gr.Dropdown(label="Scheduler (Sampler)", choices=schedulers, value="Euler a")
-                    with gr.Row():
-                        with gr.Column():
-                            with gr.Row():
-                                sdxl_lora1 = gr.Dropdown(label="LoRA1", choices=sdxl_loras, value="", allow_custom_value=True, min_width=320)
+                        sdxl_clip = gr.Dropdown(label="CLIP", choices=clips, value="", allow_custom_value=True)
+                    with gr.Column():
+                        with gr.Row():
+                            sdxl_lora1 = gr.Dropdown(label="LoRA1", choices=sdxl_loras, value="", allow_custom_value=True, min_width=320, scale=2)
                             sdxl_lora1s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA1 weight scale")
-                        with gr.Column():
-                            with gr.Row():
-                                sdxl_lora2 = gr.Dropdown(label="LoRA2", choices=sdxl_loras, value="", allow_custom_value=True, min_width=320)
+                        with gr.Row():
+                            sdxl_lora2 = gr.Dropdown(label="LoRA2", choices=sdxl_loras, value="", allow_custom_value=True, min_width=320, scale=2)
                             sdxl_lora2s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA2 weight scale")
-                        with gr.Column():
-                            with gr.Row():
-                                sdxl_lora3 = gr.Dropdown(label="LoRA3", choices=sdxl_loras, value="", allow_custom_value=True, min_width=320)
+                        with gr.Row():
+                            sdxl_lora3 = gr.Dropdown(label="LoRA3", choices=sdxl_loras, value="", allow_custom_value=True, min_width=320, scale=2)
                             sdxl_lora3s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA3 weight scale")
-                        with gr.Column():
-                            with gr.Row():
-                                sdxl_lora4 = gr.Dropdown(label="LoRA4", choices=sdxl_loras, value="", allow_custom_value=True, min_width=320)
+                        with gr.Row():
+                            sdxl_lora4 = gr.Dropdown(label="LoRA4", choices=sdxl_loras, value="", allow_custom_value=True, min_width=320, scale=2)
                             sdxl_lora4s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA4 weight scale")
-                        with gr.Column():
-                            with gr.Row():
-                                sdxl_lora5 = gr.Dropdown(label="LoRA5", choices=sdxl_loras, value="", allow_custom_value=True, min_width=320)
+                        with gr.Row():
+                            sdxl_lora5 = gr.Dropdown(label="LoRA5", choices=sdxl_loras, value="", allow_custom_value=True, min_width=320, scale=2)
                             sdxl_lora5s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA5 weight scale")
             sdxl_run_button = gr.Button(value="Submit", variant="primary")
         with gr.Tab("SD 1.5"):
             with gr.Group():
-                with gr.Row():
-                    sd15_presets = gr.Radio(label="Presets", choices=list(sd15_preset_dict.keys()), value=list(sd15_preset_dict.keys())[0])
-                with gr.Row():
-                    sd15_mtype = gr.Textbox(value="SD 1.5", visible=False)
-                with gr.Row():
-                    sd15_dtype = gr.Radio(label="Output data type", choices=get_dtypes(), value=DEFAULT_DTYPE)
+                sd15_presets = gr.Radio(label="Presets", choices=list(sd15_preset_dict.keys()), value=list(sd15_preset_dict.keys())[0])
+                sd15_mtype = gr.Textbox(value="SD 1.5", visible=False)
+                sd15_dtype = gr.Radio(label="Output data type", choices=get_dtypes(), value=DEFAULT_DTYPE)
                 with gr.Row():
                     sd15_ema = gr.Checkbox(label="Extract EMA", value=True, visible=True)
-                with gr.Row():
                     sd15_isize = gr.Radio(label="Image size", choices=["768", "512"], value="768")
-                with gr.Row():
                     sd15_sc = gr.Checkbox(label="Safety checker", value=False)
                 with gr.Accordion("Advanced settings", open=False):
                     with gr.Row():
                         sd15_vae = gr.Dropdown(label="VAE", choices=sd15_vaes, value="", allow_custom_value=True)
-                    with gr.Row():
-                        sd15_clip = gr.Dropdown(label="CLIP", choices=clips, value="", allow_custom_value=True)
-                    with gr.Row():
                         sd15_scheduler = gr.Dropdown(label="Scheduler (Sampler)", choices=schedulers, value="Euler")
-                    with gr.Row():
-                        with gr.Column():
-                            with gr.Row():
-                                sd15_lora1 = gr.Dropdown(label="LoRA1", choices=sd15_loras, value="", allow_custom_value=True, min_width=320)
-                            with gr.Row():
-                                sd15_lora1s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA1 weight scale")
-                            with gr.Row():
-                                with gr.Column():
-                                    sd15_lora2 = gr.Dropdown(label="LoRA2", choices=sd15_loras, value="", allow_custom_value=True, min_width=320)
-                                with gr.Row():
-                                    sd15_lora2s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA2 weight scale")
-                        with gr.Column():
-                            with gr.Row():
-                                sd15_lora3 = gr.Dropdown(label="LoRA3", choices=sd15_loras, value="", allow_custom_value=True, min_width=320)
-                            with gr.Row():
-                                sd15_lora3s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA3 weight scale")
-                        
-                        with gr.Column():
-                            with gr.Row():
-                                sd15_lora4 = gr.Dropdown(label="LoRA4", choices=sd15_loras, value="", allow_custom_value=True, min_width=320)
-                            with gr.Row():
-                                sd15_lora4s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA4 weight scale")
-                        with gr.Column():
-                            with gr.Row():
-                                sd15_lora5 = gr.Dropdown(label="LoRA5", choices=sd15_loras, value="", allow_custom_value=True, min_width=320)
-                            with gr.Row():
-                                sd15_lora5s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA5 weight scale")
+                        sd15_clip = gr.Dropdown(label="CLIP", choices=clips, value="", allow_custom_value=True)
+                    with gr.Column():
+                        with gr.Row():
+                            sd15_lora1 = gr.Dropdown(label="LoRA1", choices=sd15_loras, value="", allow_custom_value=True, min_width=320, scale=2)
+                            sd15_lora1s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA1 weight scale")
+                        with gr.Row():
+                            sd15_lora2 = gr.Dropdown(label="LoRA2", choices=sd15_loras, value="", allow_custom_value=True, min_width=320, scale=2)
+                            sd15_lora2s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA2 weight scale")
+                        with gr.Row():
+                            sd15_lora3 = gr.Dropdown(label="LoRA3", choices=sd15_loras, value="", allow_custom_value=True, min_width=320, scale=2)
+                            sd15_lora3s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA3 weight scale")
+                        with gr.Row():
+                            sd15_lora4 = gr.Dropdown(label="LoRA4", choices=sd15_loras, value="", allow_custom_value=True, min_width=320, scale=2)
+                            sd15_lora4s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA4 weight scale")
+                        with gr.Row():
+                            sd15_lora5 = gr.Dropdown(label="LoRA5", choices=sd15_loras, value="", allow_custom_value=True, min_width=320, scale=2)
+                            sd15_lora5s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA5 weight scale")
             sd15_run_button = gr.Button(value="Submit", variant="primary")
         with gr.Tab("FLUX.1"):
             with gr.Group():
                 flux_presets = gr.Radio(label="Presets", choices=list(flux_preset_dict.keys()), value=list(flux_preset_dict.keys())[0])
                 flux_mtype = gr.Textbox(value="FLUX", visible=False)
-                with gr.Row():
-                    flux_dtype = gr.Radio(label="Output data type", choices=get_dtypes(), value="bf16")
+                flux_dtype = gr.Radio(label="Output data type", choices=get_dtypes(), value="bf16")
                 flux_base_repo = gr.Dropdown(label="Base repo ID", choices=FLUX_BASE_REPOS, value=FLUX_BASE_REPOS[0], allow_custom_value=True, visible=True)
                 with gr.Accordion("Advanced settings", open=False):
                     with gr.Row():
                         flux_vae = gr.Dropdown(label="VAE", choices=flux_vaes, value="", allow_custom_value=True)
+                        flux_scheduler = gr.Dropdown(label="Scheduler (Sampler)", choices=[""], value="", visible=False)
                     with gr.Row():
                         flux_clip = gr.Dropdown(label="CLIP", choices=clips, value="", allow_custom_value=True)
-                    with gr.Row():
                         flux_t5 = gr.Dropdown(label="T5", choices=t5s, value="", allow_custom_value=True)
-                        flux_scheduler = gr.Dropdown(label="Scheduler (Sampler)", choices=[""], value="", visible=False)
-                    with gr.Row():
-                        with gr.Column():
-                            flux_lora1 = gr.Dropdown(label="LoRA1", choices=flux_loras, value="", allow_custom_value=True, min_width=320)
+                    with gr.Column():
+                        with gr.Row():
+                            flux_lora1 = gr.Dropdown(label="LoRA1", choices=flux_loras, value="", allow_custom_value=True, min_width=320, scale=2)
                             flux_lora1s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA1 weight scale")
-                        with gr.Column():
-                            flux_lora2 = gr.Dropdown(label="LoRA2", choices=flux_loras, value="", allow_custom_value=True, min_width=320)
+                        with gr.Row():
+                            flux_lora2 = gr.Dropdown(label="LoRA2", choices=flux_loras, value="", allow_custom_value=True, min_width=320, scale=2)
                             flux_lora2s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA2 weight scale")
-                        with gr.Column():
-                            flux_lora3 = gr.Dropdown(label="LoRA3", choices=flux_loras, value="", allow_custom_value=True, min_width=320)
+                        with gr.Row():
+                            flux_lora3 = gr.Dropdown(label="LoRA3", choices=flux_loras, value="", allow_custom_value=True, min_width=320, scale=2)
                             flux_lora3s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA3 weight scale")
-                        with gr.Column():
-                            flux_lora4 = gr.Dropdown(label="LoRA4", choices=flux_loras, value="", allow_custom_value=True, min_width=320)
+                        with gr.Row():
+                            flux_lora4 = gr.Dropdown(label="LoRA4", choices=flux_loras, value="", allow_custom_value=True, min_width=320, scale=2)
                             flux_lora4s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA4 weight scale")
-                        with gr.Column():
-                            flux_lora5 = gr.Dropdown(label="LoRA5", choices=flux_loras, value="", allow_custom_value=True, min_width=320)
+                        with gr.Row():
+                            flux_lora5 = gr.Dropdown(label="LoRA5", choices=flux_loras, value="", allow_custom_value=True, min_width=320, scale=2)
                             flux_lora5s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA5 weight scale")
             flux_run_button = gr.Button(value="Submit", variant="primary")
         with gr.Tab("SD 3.5"):
             with gr.Group():
-                with gr.Row():
-                    sd35_presets = gr.Radio(label="Presets", choices=list(sd35_preset_dict.keys()), value=list(sd35_preset_dict.keys())[0])
-                with gr.Row():
-                    sd35_mtype = gr.Textbox(value="SD 3.5", visible=False)
-                with gr.Row():
-                    sd35_dtype = gr.Radio(label="Output data type", choices=get_dtypes(), value="bf16")
+                sd35_presets = gr.Radio(label="Presets", choices=list(sd35_preset_dict.keys()), value=list(sd35_preset_dict.keys())[0])
+                sd35_mtype = gr.Textbox(value="SD 3.5", visible=False)
+                sd35_dtype = gr.Radio(label="Output data type", choices=get_dtypes(), value="bf16")
                 sd35_base_repo = gr.Dropdown(label="Base repo ID", choices=SD35_BASE_REPOS, value=SD35_BASE_REPOS[0], allow_custom_value=True, visible=True)
                 with gr.Accordion("Advanced settings", open=False):
                     with gr.Row():
                         sd35_vae = gr.Dropdown(label="VAE", choices=sd35_vaes, value="", allow_custom_value=True)
-                        sd35_clip = gr.Dropdown(label="CLIP", choices=clips, value="", allow_custom_value=True)
-                        sd35_t5 = gr.Dropdown(label="T5", choices=t5s, value="", allow_custom_value=True)
                         sd35_scheduler = gr.Dropdown(label="Scheduler (Sampler)", choices=[""], value="", visible=False)
                     with gr.Row():
-                        with gr.Column():
-                            with gr.Row():
-                                sd35_lora1 = gr.Dropdown(label="LoRA1", choices=sd35_loras, value="", allow_custom_value=True, min_width=320)
-                                sd35_lora1s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA1 weight scale")
-                        with gr.Column():
-                            with gr.Row():
-                                sd35_lora2 = gr.Dropdown(label="LoRA2", choices=sd35_loras, value="", allow_custom_value=True, min_width=320)
-                                sd35_lora2s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA2 weight scale")
-                        with gr.Column():
-                            with gr.Row():
-                                sd35_lora3 = gr.Dropdown(label="LoRA3", choices=sd35_loras, value="", allow_custom_value=True, min_width=320)
-                                sd35_lora3s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA3 weight scale")
-                        with gr.Column():
-                            with gr.Row():
-                                sd35_lora4 = gr.Dropdown(label="LoRA4", choices=sd35_loras, value="", allow_custom_value=True, min_width=320)
-                                sd35_lora4s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA4 weight scale")
-                        with gr.Column():
-                            with gr.Row():
-                                sd35_lora5 = gr.Dropdown(label="LoRA5", choices=sd35_loras, value="", allow_custom_value=True, min_width=320)
-                                sd35_lora5s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA5 weight scale")
+                        sd35_clip = gr.Dropdown(label="CLIP", choices=clips, value="", allow_custom_value=True)
+                        sd35_t5 = gr.Dropdown(label="T5", choices=t5s, value="", allow_custom_value=True)
+                    with gr.Column():
+                        with gr.Row():
+                            sd35_lora1 = gr.Dropdown(label="LoRA1", choices=sd35_loras, value="", allow_custom_value=True, min_width=320, scale=2)
+                            sd35_lora1s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA1 weight scale")
+                        with gr.Row():
+                            sd35_lora2 = gr.Dropdown(label="LoRA2", choices=sd35_loras, value="", allow_custom_value=True, min_width=320, scale=2)
+                            sd35_lora2s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA2 weight scale")
+                        with gr.Row():
+                            sd35_lora3 = gr.Dropdown(label="LoRA3", choices=sd35_loras, value="", allow_custom_value=True, min_width=320, scale=2)
+                            sd35_lora3s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA3 weight scale")
+                        with gr.Row():
+                            sd35_lora4 = gr.Dropdown(label="LoRA4", choices=sd35_loras, value="", allow_custom_value=True, min_width=320, scale=2)
+                            sd35_lora4s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA4 weight scale")
+                        with gr.Row():
+                            sd35_lora5 = gr.Dropdown(label="LoRA5", choices=sd35_loras, value="", allow_custom_value=True, min_width=320, scale=2)
+                            sd35_lora5s = gr.Slider(minimum=-2, maximum=2, step=0.01, value=1.00, label="LoRA5 weight scale")
             sd35_run_button = gr.Button(value="Submit", variant="primary")
         adv_args = gr.Textbox(label="Advanced arguments", value="", visible=False)
         with gr.Group():
@@ -281,4 +248,4 @@ In particular, conversion of FLUX.1 or SD 3.5 is almost impossible in CPU space.
     clear_button.click(lambda: ([], "<br><br>"), None, [repo_urls, output_md], queue=False, show_api=False)
 
 demo.queue()
-demo.launch()
+demo.launch(ssr_mode=False)

presets.py

@@ -1,134 +1,147 @@
-from sdutils import get_dtypes, SCHEDULER_CONFIG_MAP
-
-
-DEFAULT_DTYPE = get_dtypes()[0]
-schedulers = list(SCHEDULER_CONFIG_MAP.keys())
-
-
-clips = [
-    "",
-    "openai/clip-vit-large-patch14",
-]
-
-
-t5s = [
-    "",
-    "https://huggingface.co/camenduru/FLUX.1-dev/blob/main/t5xxl_fp8_e4m3fn.safetensors",
-]
-
-
-sdxl_vaes = [
-    "",
-    "https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl.vae.safetensors",
-    "https://huggingface.co/nubby/blessed-sdxl-vae-fp16-fix/blob/main/sdxl_vae-fp16fix-blessed.safetensors",
-    "https://huggingface.co/John6666/safetensors_converting_test/blob/main/xlVAEC_e7.safetensors",
-    "https://huggingface.co/John6666/safetensors_converting_test/blob/main/xlVAEC_f1.safetensors",
-]
-
-
-sdxl_loras = [
-    "",
-    "https://huggingface.co/SPO-Diffusion-Models/SPO-SDXL_4k-p_10ep_LoRA/blob/main/spo_sdxl_10ep_4k-data_lora_diffusers.safetensors",
-    "https://huggingface.co/wangfuyun/PCM_Weights/blob/main/sdxl/pcm_sdxl_smallcfg_2step_converted.safetensors",
-    "https://huggingface.co/wangfuyun/PCM_Weights/blob/main/sdxl/pcm_sdxl_smallcfg_4step_converted.safetensors",
-    "https://huggingface.co/wangfuyun/PCM_Weights/blob/main/sdxl/pcm_sdxl_smallcfg_8step_converted.safetensors",
-    "https://huggingface.co/wangfuyun/PCM_Weights/blob/main/sdxl/pcm_sdxl_normalcfg_8step_converted.safetensors",
-    "https://huggingface.co/wangfuyun/PCM_Weights/blob/main/sdxl/pcm_sdxl_normalcfg_16step_converted.safetensors",
-    "https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-1step-lora.safetensors",
-    "https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-2steps-lora.safetensors",
-    "https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-4steps-lora.safetensors",
-    "https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-8steps-CFG-lora.safetensors",
-    "https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-12steps-CFG-lora.safetensors",
-    "https://huggingface.co/latent-consistency/lcm-lora-sdxl/blob/main/pytorch_lora_weights.safetensors",
-]
-
-
-sdxl_preset_dict = {
-    "Default": [DEFAULT_DTYPE, "", "Euler a", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0],
-    "Bake in standard VAE": [DEFAULT_DTYPE, "https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl.vae.safetensors",
-                             "Euler a", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0],
-    "Hyper-SDXL / SPO": [DEFAULT_DTYPE, "https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl.vae.safetensors",
-     "TCD", "https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-8steps-CFG-lora.safetensors", 1.0,
-     "https://huggingface.co/SPO-Diffusion-Models/SPO-SDXL_4k-p_10ep_LoRA/blob/main/spo_sdxl_10ep_4k-data_lora_diffusers.safetensors",
-       1.0, "", 1.0, "", 1.0, "", 1.0],
-}
-
-
-def sdxl_set_presets(preset: str="Default"):
-    p = []
-    if preset in sdxl_preset_dict.keys(): p = sdxl_preset_dict[preset]
-    else: p = sdxl_preset_dict["Default"]
-    return p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], p[8], p[9], p[10], p[11], p[12]
-
-
-sd15_vaes = [
-    "",
-    "https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt",
-    "https://huggingface.co/stabilityai/sd-vae-ft-ema-original/resolve/main/vae-ft-ema-560000-ema-pruned.ckpt",
-]
-
-
-sd15_loras = [
-    "",
-    "https://huggingface.co/SPO-Diffusion-Models/SPO-SD-v1-5_4k-p_10ep_LoRA/blob/main/spo-sd-v1-5_4k-p_10ep_lora_diffusers.safetensors",
-]
-
-
-sd15_preset_dict = {
-    "Default": [DEFAULT_DTYPE, "", "Euler", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0, True],
-    "Bake in standard VAE": [DEFAULT_DTYPE, "https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt",
-                             "Euler", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0, True],
-}
-
-
-def sd15_set_presets(preset: str="Default"):
-    p = []
-    if preset in sd15_preset_dict.keys(): p = sd15_preset_dict[preset]
-    else: p = sd15_preset_dict["Default"]
-    return p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], p[8], p[9], p[10], p[11], p[12], p[13]
-
-
-flux_vaes = [
-    "",
-]
-
-
-flux_loras = [
-    "",
-]
-
-
-flux_preset_dict = {
-    "dev": ["bf16", "", "", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0, "camenduru/FLUX.1-dev-diffusers"],
-    "schnell": ["bf16", "", "", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0, "black-forest-labs/FLUX.1-schnell"],
-}
-
-
-def flux_set_presets(preset: str="dev"):
-    p = []
-    if preset in flux_preset_dict.keys(): p = flux_preset_dict[preset]
-    else: p = flux_preset_dict["dev"]
-    return p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], p[8], p[9], p[10], p[11], p[12], p[13]
-
-
-
-sd35_vaes = [
-    "",
-]
-
-
-sd35_loras = [
-    "",
-]
-
-
-sd35_preset_dict = {
-    "Default": ["bf16", "", "", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0, "adamo1139/stable-diffusion-3.5-large-ungated"],
-}
-
-
-def sd35_set_presets(preset: str="dev"):
-    p = []
-    if preset in sd35_preset_dict.keys(): p = sd35_preset_dict[preset]
-    else: p = sd35_preset_dict["Default"]
-    return p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], p[8], p[9], p[10], p[11], p[12], p[13]
+from sdutils import get_dtypes, SCHEDULER_CONFIG_MAP
+import gradio as gr
+
+
+DEFAULT_DTYPE = get_dtypes()[0]
+schedulers = list(SCHEDULER_CONFIG_MAP.keys())
+
+
+clips = [
+    "",
+    "openai/clip-vit-large-patch14",
+]
+
+
+t5s = [
+    "",
+    "https://huggingface.co/camenduru/FLUX.1-dev/blob/main/t5xxl_fp8_e4m3fn.safetensors",
+]
+
+
+sdxl_vaes = [
+    "",
+    "https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl.vae.safetensors",
+    "https://huggingface.co/nubby/blessed-sdxl-vae-fp16-fix/blob/main/sdxl_vae-fp16fix-blessed.safetensors",
+    "https://huggingface.co/John6666/safetensors_converting_test/blob/main/xlVAEC_e7.safetensors",
+    "https://huggingface.co/John6666/safetensors_converting_test/blob/main/xlVAEC_f1.safetensors",
+]
+
+
+sdxl_loras = [
+    "",
+    "https://huggingface.co/SPO-Diffusion-Models/SPO-SDXL_4k-p_10ep_LoRA/blob/main/spo_sdxl_10ep_4k-data_lora_diffusers.safetensors",
+    "https://huggingface.co/wangfuyun/PCM_Weights/blob/main/sdxl/pcm_sdxl_smallcfg_2step_converted.safetensors",
+    "https://huggingface.co/wangfuyun/PCM_Weights/blob/main/sdxl/pcm_sdxl_smallcfg_4step_converted.safetensors",
+    "https://huggingface.co/wangfuyun/PCM_Weights/blob/main/sdxl/pcm_sdxl_smallcfg_8step_converted.safetensors",
+    "https://huggingface.co/wangfuyun/PCM_Weights/blob/main/sdxl/pcm_sdxl_normalcfg_8step_converted.safetensors",
+    "https://huggingface.co/wangfuyun/PCM_Weights/blob/main/sdxl/pcm_sdxl_normalcfg_16step_converted.safetensors",
+    "https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-1step-lora.safetensors",
+    "https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-2steps-lora.safetensors",
+    "https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-4steps-lora.safetensors",
+    "https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-8steps-CFG-lora.safetensors",
+    "https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-12steps-CFG-lora.safetensors",
+    "https://huggingface.co/latent-consistency/lcm-lora-sdxl/blob/main/pytorch_lora_weights.safetensors",
+]
+
+
+sdxl_preset_items = ["dtype", "vae", "scheduler", "lora1", "lora1s", "lora2", "lora2s", "lora3", "lora3s", "lora4", "lora4s", "lora5", "lora5s"]
+sdxl_preset_dict = {
+    "Default": [DEFAULT_DTYPE, "", "Euler a", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0],
+    "Bake in standard VAE": [DEFAULT_DTYPE, "https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl.vae.safetensors",
+                             "Euler a", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0],
+    "Hyper-SDXL / SPO": [DEFAULT_DTYPE, "https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl.vae.safetensors",
+     "TCD", "https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-8steps-CFG-lora.safetensors", 1.0,
+     "https://huggingface.co/SPO-Diffusion-Models/SPO-SDXL_4k-p_10ep_LoRA/blob/main/spo_sdxl_10ep_4k-data_lora_diffusers.safetensors",
+       1.0, "", 1.0, "", 1.0, "", 1.0],
+}
+
+
+def sdxl_set_presets(preset: str="Default"):
+    p = []
+    if preset in sdxl_preset_dict.keys(): p = sdxl_preset_dict[preset]
+    else: p = sdxl_preset_dict["Default"]
+    if len(p) != len(sdxl_preset_items): raise gr.Error("Invalid preset.")
+    print("Setting SDXL preset:", ", ".join([f"{x}:{y}" for x, y in zip(sdxl_preset_items, p)]))
+    return p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], p[8], p[9], p[10], p[11], p[12]
+
+
+sd15_vaes = [
+    "",
+    "https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt",
+    "https://huggingface.co/stabilityai/sd-vae-ft-ema-original/resolve/main/vae-ft-ema-560000-ema-pruned.ckpt",
+]
+
+
+sd15_loras = [
+    "",
+    "https://huggingface.co/SPO-Diffusion-Models/SPO-SD-v1-5_4k-p_10ep_LoRA/blob/main/spo-sd-v1-5_4k-p_10ep_lora_diffusers.safetensors",
+]
+
+
+sd15_preset_items = ["dtype", "vae", "scheduler", "lora1", "lora1s", "lora2", "lora2s", "lora3", "lora3s", "lora4", "lora4s", "lora5", "lora5s", "ema"]
+sd15_preset_dict = {
+    "Default": [DEFAULT_DTYPE, "", "Euler", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0, True],
+    "Bake in standard VAE": [DEFAULT_DTYPE, "https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt",
+                             "Euler", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0, True],
+}
+
+
+def sd15_set_presets(preset: str="Default"):
+    p = []
+    if preset in sd15_preset_dict.keys(): p = sd15_preset_dict[preset]
+    else: p = sd15_preset_dict["Default"]
+    if len(p) != len(sd15_preset_items): raise gr.Error("Invalid preset.")
+    print("Setting SD1.5 preset:", ", ".join([f"{x}:{y}" for x, y in zip(sd15_preset_items, p)]))
+    return p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], p[8], p[9], p[10], p[11], p[12], p[13]
+
+
+flux_vaes = [
+    "",
+]
+
+
+flux_loras = [
+    "",
+]
+
+
+flux_preset_items = ["dtype", "vae", "scheduler", "lora1", "lora1s", "lora2", "lora2s", "lora3", "lora3s", "lora4", "lora4s", "lora5", "lora5s", "base_repo"]
+flux_preset_dict = {
+    "dev": ["bf16", "", "", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0, "camenduru/FLUX.1-dev-diffusers"],
+    "schnell": ["bf16", "", "", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0, "black-forest-labs/FLUX.1-schnell"],
+}
+
+
+def flux_set_presets(preset: str="dev"):
+    p = []
+    if preset in flux_preset_dict.keys(): p = flux_preset_dict[preset]
+    else: p = flux_preset_dict["dev"]
+    if len(p) != len(flux_preset_items): raise gr.Error("Invalid preset.")
+    print("Setting FLUX.1 preset:", ", ".join([f"{x}:{y}" for x, y in zip(flux_preset_items, p)]))
+    return p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], p[8], p[9], p[10], p[11], p[12], p[13]
+
+
+
+sd35_vaes = [
+    "",
+]
+
+
+sd35_loras = [
+    "",
+]
+
+
+sd35_preset_items = ["dtype", "vae", "scheduler", "lora1", "lora1s", "lora2", "lora2s", "lora3", "lora3s", "lora4", "lora4s", "lora5", "lora5s", "base_repo"]
+sd35_preset_dict = {
+    "Default": ["bf16", "", "", "", 1.0, "", 1.0, "", 1.0, "", 1.0, "", 1.0, "adamo1139/stable-diffusion-3.5-large-ungated"],
+}
+
+
+def sd35_set_presets(preset: str="dev"):
+    p = []
+    if preset in sd35_preset_dict.keys(): p = sd35_preset_dict[preset]
+    else: p = sd35_preset_dict["Default"]
+    if len(p) != len(sd35_preset_items): raise gr.Error("Invalid preset.")
+    print("Setting SD3.5 preset:", ", ".join([f"{x}:{y}" for x, y in zip(sd35_preset_items, p)]))
+    return p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], p[8], p[9], p[10], p[11], p[12], p[13]