v5 release with triple architecture support and prompt enhancer
This commit is contained in:
17
ltx_video/configs/ltxv-2b-0.9.6-dev.yaml
Normal file
17
ltx_video/configs/ltxv-2b-0.9.6-dev.yaml
Normal file
@@ -0,0 +1,17 @@
|
||||
pipeline_type: base
|
||||
checkpoint_path: "ltxv-2b-0.9.6-dev-04-25.safetensors"
|
||||
guidance_scale: 3
|
||||
stg_scale: 1
|
||||
rescaling_scale: 0.7
|
||||
skip_block_list: [19]
|
||||
num_inference_steps: 40
|
||||
stg_mode: "attention_values" # options: "attention_values", "attention_skip", "residual", "transformer_block"
|
||||
decode_timestep: 0.05
|
||||
decode_noise_scale: 0.025
|
||||
text_encoder_model_name_or_path: "PixArt-alpha/PixArt-XL-2-1024-MS"
|
||||
precision: "bfloat16"
|
||||
sampler: "from_checkpoint" # options: "uniform", "linear-quadratic", "from_checkpoint"
|
||||
prompt_enhancement_words_threshold: 120
|
||||
prompt_enhancer_image_caption_model_name_or_path: "MiaoshouAI/Florence-2-large-PromptGen-v2.0"
|
||||
prompt_enhancer_llm_model_name_or_path: "unsloth/Llama-3.2-3B-Instruct"
|
||||
stochastic_sampling: false
|
||||
Reference in New Issue
Block a user