Davids048andClaude Fable 5.1
a5aa64ab6b
VideoGenerator.from_pretrained(model_path, config) takes the same nested
settings as from_config; the 24 flat keywords, from_pretrained_kwargs_to_config,
FROM_PRETRAINED_KWARGS, and every flat_field in the schema are deleted.
nvfp4_fa4 is the typed field engine.attention.nvfp4_fa4, applied by a
resolution step. 65 call sites and the docs use the nested form; the three
kwargs golden cases are config cases with byte-identical results.
Every value is decided during resolution; no runtime code calls
with_override. The device offload policy (unified memory, MPS, layerwise
conflicts, lazy module load) runs as resolution steps in the main process
(fastvideo/api/device_policy.py), and the worker runs on the config it
receives. The LTX-2 refine defaults from model_index.json and the MiniMax-H3
schedule from fastvideo_inference.json are resolution steps
(fastvideo/api/checkpoint_defaults.py) that read local checkpoints or download
only those files; the H3 pipeline validates the schedule against the loaded
schedulers but no longer writes it. The preprocessing entry point passes the
downloaded local path as run state instead of a config override. Teacher and
critic models load with override_transformer_cls_name as a load argument.
The test isolation turns the device policy off, as it blocks downloads, so
that the goldens do not depend on the machine.
Against the 3f6893a0 goldens, every field still matches except boundary_ratio
and ltx2_vae_tiling (DESIGN section 9) and the fields that no longer exist.
The launcher and trainer YAML verifications report 0 unexplained differences.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
413 lines
13 KiB
Python
413 lines
13 KiB
Python
# SPDX-License-Identifier: Apache-2.0
|
|
import json
|
|
|
|
import yaml
|
|
|
|
from fastvideo.api import (
|
|
config_to_dict,
|
|
ContinuationState,
|
|
GenerationRequest,
|
|
GeneratorConfig,
|
|
GpuPoolConfig,
|
|
load_run_config,
|
|
load_serve_config,
|
|
parse_config,
|
|
PlannedStage,
|
|
PromptEnhancerConfig,
|
|
PromptSafetyConfig,
|
|
RunConfig,
|
|
ServeConfig,
|
|
StreamingConfig,
|
|
WarmupConfig,
|
|
)
|
|
|
|
|
|
def test_parse_config_builds_nested_typed_config() -> None:
|
|
raw = {
|
|
"generator": {
|
|
"model_path": "/models/ltx2",
|
|
"pipeline": {
|
|
"workload_type": "t2v",
|
|
"preset": "ltx2_two_stage",
|
|
},
|
|
},
|
|
"request": {
|
|
"prompt": ["a fox", "a wolf"],
|
|
"sampling": {
|
|
"num_frames": 121,
|
|
"height": 1024,
|
|
"width": 1536,
|
|
"guidance_scale": 1.5,
|
|
},
|
|
"state": {
|
|
"kind": "ltx2_continuation",
|
|
"payload": {
|
|
"segment_index": 1
|
|
},
|
|
},
|
|
"plan": {
|
|
"stages": [{
|
|
"name": "base",
|
|
"kind": "sample",
|
|
}]
|
|
},
|
|
},
|
|
}
|
|
|
|
config = parse_config(RunConfig, raw)
|
|
|
|
assert config.generator.pipeline.preset == "ltx2_two_stage"
|
|
assert config.request.prompt == ["a fox", "a wolf"]
|
|
assert config.request.state == ContinuationState(
|
|
kind="ltx2_continuation",
|
|
payload={"segment_index": 1},
|
|
)
|
|
assert config.request.plan is not None
|
|
assert config.request.plan.stages == [PlannedStage(name="base", kind="sample")]
|
|
|
|
|
|
def test_parse_config_accepts_existing_typed_instance() -> None:
|
|
typed = RunConfig(
|
|
generator=GeneratorConfig(model_path="/models/base"),
|
|
request=GenerationRequest(prompt="hello"),
|
|
)
|
|
|
|
assert parse_config(RunConfig, typed) is typed
|
|
|
|
|
|
def test_load_run_config_supports_yaml_roundtrip(tmp_path) -> None:
|
|
raw = {
|
|
"generator": {
|
|
"model_path": "/models/wan"
|
|
},
|
|
"request": {
|
|
"prompt": "hello",
|
|
"sampling": {
|
|
"num_frames": 16
|
|
},
|
|
},
|
|
}
|
|
path = tmp_path / "run.yaml"
|
|
path.write_text(yaml.safe_dump(raw), encoding="utf-8")
|
|
|
|
loaded = load_run_config(path)
|
|
|
|
assert config_to_dict(loaded) == {
|
|
"generator": {
|
|
"model_path": "/models/wan",
|
|
"mode": "inference",
|
|
"revision": None,
|
|
"trust_remote_code": False,
|
|
"engine": {
|
|
"num_gpus": 1,
|
|
"execution_backend": "mp",
|
|
"parallelism": {
|
|
"tp_size": -1,
|
|
"sp_size": -1,
|
|
"hsdp_replicate_dim": 1,
|
|
"hsdp_shard_dim": -1,
|
|
"dist_timeout": None,
|
|
"master_port": None,
|
|
},
|
|
"offload": {
|
|
"dit": True,
|
|
"dit_layerwise": True,
|
|
"text_encoder": True,
|
|
"image_encoder": True,
|
|
"vae": True,
|
|
"pin_cpu_memory": True,
|
|
"lazy_module_load": None,
|
|
},
|
|
"compile": {
|
|
"enabled": False,
|
|
"backend": None,
|
|
"fullgraph": None,
|
|
"mode": None,
|
|
"dynamic": None,
|
|
"extras": {},
|
|
"text_encoder_enabled": None,
|
|
"vae_enabled": None,
|
|
"audio_vae_enabled": None,
|
|
"regional": None,
|
|
"dit_kwargs": {},
|
|
"text_encoder_kwargs": {},
|
|
"vae_kwargs": {},
|
|
"audio_vae_kwargs": {},
|
|
},
|
|
"attention": {
|
|
"backend": None,
|
|
"vsa_sparsity": None,
|
|
"vsa_tile_size": None,
|
|
"moba_config_path": None,
|
|
"moba_config": None,
|
|
"nvfp4_fa4": False,
|
|
},
|
|
"precision": {
|
|
"dit": None,
|
|
"vae": None,
|
|
"vae_decode": None,
|
|
"image_encoder": None,
|
|
"text_encoders": None,
|
|
},
|
|
"enable_stage_verification": True,
|
|
"use_fsdp_inference": False,
|
|
"disable_autocast": False,
|
|
"quantization": None,
|
|
},
|
|
"pipeline": {
|
|
"workload_type": None,
|
|
"preset": None,
|
|
"preset_version": None,
|
|
"components": {
|
|
"config_root": None,
|
|
"pipeline_config_path": None,
|
|
"text_encoder_weights": None,
|
|
"transformer_weights": None,
|
|
"transformer_2_weights": None,
|
|
"vae_weights": None,
|
|
"upsampler_weights": None,
|
|
"lora_path": None,
|
|
"lora_nickname": "default",
|
|
"lora_strength": 1.0,
|
|
"lora_target_modules": None,
|
|
"override_pipeline_cls_name": None,
|
|
"override_transformer_cls_name": None,
|
|
},
|
|
"vae_tiling": None,
|
|
"vae_sp": None,
|
|
"flow_shift": None,
|
|
"embedded_cfg_scale": None,
|
|
"dmd_denoising_steps": None,
|
|
"boundary_ratio": None,
|
|
"output_type": "pil",
|
|
"dit": {},
|
|
"vae": {},
|
|
"ltx2": {
|
|
"vae_spatial_tile_size_in_pixels": None,
|
|
"vae_spatial_tile_overlap_in_pixels": None,
|
|
"vae_temporal_tile_size_in_frames": None,
|
|
"vae_temporal_tile_overlap_in_frames": None,
|
|
"initial_latent_path": None,
|
|
"audio_latent_path": None,
|
|
"legacy_native_noise_order": None,
|
|
"use_distilled_sigmas": None,
|
|
"refine": {
|
|
"enabled": None,
|
|
"num_inference_steps": None,
|
|
"guidance_scale": None,
|
|
"add_noise": None,
|
|
"image_crf": None,
|
|
"video_position_offset_sec": None,
|
|
"transformer_path": None,
|
|
"lora_path": None,
|
|
"noise_path": None,
|
|
"audio_noise_path": None,
|
|
},
|
|
},
|
|
"minimax_h3": {
|
|
"sequential_load": None,
|
|
"video_decode_backend": None,
|
|
"taeh3_checkpoint": None,
|
|
"taeh3_chunk_size": None,
|
|
"vae_parallel_decode": None,
|
|
"vae_parallel_encode": None,
|
|
"vae_parallel_decode_strategy": None,
|
|
},
|
|
"longcat": {
|
|
"enable_bsa": None,
|
|
"bsa_sparsity": None,
|
|
"bsa_cdf_threshold": None,
|
|
"bsa_chunk_q": None,
|
|
"bsa_chunk_k": None,
|
|
},
|
|
"preset_overrides": {},
|
|
"experimental": {},
|
|
},
|
|
},
|
|
"request": {
|
|
"prompt": "hello",
|
|
"negative_prompt": None,
|
|
"inputs": {
|
|
"prompt_path": None,
|
|
"image_path": None,
|
|
"video_path": None,
|
|
"pil_image": None,
|
|
"last_image": None,
|
|
"references": None,
|
|
"latents": None,
|
|
"audio_latents": None,
|
|
"pose": None,
|
|
"mouse_cond": None,
|
|
"keyboard_cond": None,
|
|
"grid_sizes": None,
|
|
"c2ws_plucker_emb": None,
|
|
"action_path": None,
|
|
"refine_from": None,
|
|
"stage1_video": None,
|
|
},
|
|
"sampling": {
|
|
"num_videos_per_prompt": 1,
|
|
"seed": 1024,
|
|
"max_sequence_length": None,
|
|
"num_frames": 16,
|
|
"height": 720,
|
|
"width": 1280,
|
|
"height_sr": 1072,
|
|
"width_sr": 1920,
|
|
"fps": 24,
|
|
"num_inference_steps": 50,
|
|
"num_inference_steps_sr": 50,
|
|
"guidance_scale": 1.0,
|
|
"batch_cfg": False,
|
|
"guidance_scale_2": None,
|
|
"cfg_normalization": False,
|
|
"cfg_truncation": 1.0,
|
|
"guidance_rescale": 0.0,
|
|
"true_cfg_scale": None,
|
|
"use_embedded_guidance": None,
|
|
"embedded_cfg_scale": None,
|
|
"boundary_ratio": None,
|
|
"sigmas": None,
|
|
},
|
|
"runtime": {
|
|
"enable_teacache": False,
|
|
"return_trajectory_latents": False,
|
|
"return_trajectory_decoded": False,
|
|
},
|
|
"output": {
|
|
"output_path": "outputs/",
|
|
"output_video_name": None,
|
|
"save_video": True,
|
|
"return_frames": True,
|
|
"return_state": False,
|
|
},
|
|
"stage_overrides": {},
|
|
"state": None,
|
|
"plan": None,
|
|
"extensions": {},
|
|
},
|
|
}
|
|
|
|
|
|
def test_load_serve_config_supports_json_roundtrip(tmp_path) -> None:
|
|
raw = {
|
|
"generator": {
|
|
"model_path": "/models/server"
|
|
},
|
|
"server": {
|
|
"port": 9000
|
|
},
|
|
"default_request": {
|
|
"prompt": "serve default"
|
|
},
|
|
}
|
|
path = tmp_path / "serve.json"
|
|
path.write_text(json.dumps(raw), encoding="utf-8")
|
|
|
|
loaded = load_serve_config(path)
|
|
|
|
assert isinstance(loaded, ServeConfig)
|
|
assert loaded.server.port == 9000
|
|
assert loaded.default_request.prompt == "serve default"
|
|
|
|
|
|
def test_serve_config_streaming_defaults_to_none() -> None:
|
|
raw = {"generator": {"model_path": "/models/server"}}
|
|
loaded = parse_config(ServeConfig, raw)
|
|
assert loaded.streaming is None
|
|
|
|
|
|
def test_serve_config_parses_streaming_block() -> None:
|
|
raw = {
|
|
"generator": {
|
|
"model_path": "/models/server"
|
|
},
|
|
"streaming": {
|
|
"session_timeout_seconds": 120,
|
|
"generation_segment_cap": 4,
|
|
"stream_mode": "legacy_jpeg",
|
|
"warmup": {
|
|
"enabled": False,
|
|
"prompt": "warmup prompt",
|
|
"timeout_seconds": 600,
|
|
},
|
|
"pool": {
|
|
"num_workers": 2,
|
|
"enable_audio_reencode": False,
|
|
"conditioning_num_frames": 5,
|
|
"conditioning_end_offset": 1,
|
|
},
|
|
"prompt": {
|
|
"provider": "groq",
|
|
"model": "llama-3-70b",
|
|
"timeout_ms": 10000,
|
|
"system_prompt_dir": "/opt/prompts",
|
|
},
|
|
"safety": {
|
|
"enabled": True,
|
|
"classifier_path": "/opt/safety.pt",
|
|
},
|
|
},
|
|
}
|
|
|
|
loaded = parse_config(ServeConfig, raw)
|
|
|
|
assert isinstance(loaded.streaming, StreamingConfig)
|
|
assert loaded.streaming.session_timeout_seconds == 120
|
|
assert loaded.streaming.generation_segment_cap == 4
|
|
assert loaded.streaming.stream_mode == "legacy_jpeg"
|
|
assert loaded.streaming.warmup == WarmupConfig(enabled=False, prompt="warmup prompt", timeout_seconds=600)
|
|
assert loaded.streaming.pool == GpuPoolConfig(
|
|
num_workers=2,
|
|
enable_audio_reencode=False,
|
|
conditioning_num_frames=5,
|
|
conditioning_end_offset=1,
|
|
)
|
|
assert loaded.streaming.prompt == PromptEnhancerConfig(
|
|
provider="groq",
|
|
model="llama-3-70b",
|
|
timeout_ms=10000,
|
|
system_prompt_dir="/opt/prompts",
|
|
)
|
|
assert loaded.streaming.safety == PromptSafetyConfig(enabled=True, classifier_path="/opt/safety.pt")
|
|
|
|
|
|
def test_serve_config_streaming_round_trip_through_config_to_dict() -> None:
|
|
raw = {
|
|
"generator": {
|
|
"model_path": "/models/server"
|
|
},
|
|
"streaming": {
|
|
"session_timeout_seconds": 600
|
|
},
|
|
}
|
|
loaded = parse_config(ServeConfig, raw)
|
|
dumped = config_to_dict(loaded)
|
|
assert dumped["streaming"]["session_timeout_seconds"] == 600
|
|
assert dumped["streaming"]["warmup"]["enabled"] is True
|
|
assert dumped["streaming"]["prompt"]["enabled"] is False
|
|
assert dumped["streaming"]["prompt"]["provider"] == "cerebras"
|
|
assert dumped["streaming"]["safety"]["enabled"] is False
|
|
|
|
|
|
def test_load_serve_config_with_streaming_from_yaml(tmp_path) -> None:
|
|
raw = {
|
|
"generator": {
|
|
"model_path": "/models/server"
|
|
},
|
|
"streaming": {
|
|
"stream_mode": "av_fmp4",
|
|
"pool": {
|
|
"num_workers": 4
|
|
},
|
|
},
|
|
}
|
|
path = tmp_path / "serve.yaml"
|
|
path.write_text(yaml.safe_dump(raw), encoding="utf-8")
|
|
|
|
loaded = load_serve_config(path)
|
|
|
|
assert loaded.streaming is not None
|
|
assert loaded.streaming.stream_mode == "av_fmp4"
|
|
assert loaded.streaming.pool.num_workers == 4
|