lars
c4b12b5e7a
CI / Format (ruff format) (push) Successful in 28s
CI / Lint (ruff check) (push) Successful in 29s
CI / Sync project version with tag (push) Has been skipped
CI / Lint (ruff check) (pull_request) Successful in 40s
CI / Type check (ty) (push) Successful in 45s
CI / Format (ruff format) (pull_request) Successful in 33s
CI / Sync project version with tag (pull_request) Has been skipped
CI / Type check (ty) (pull_request) Successful in 36s
CI / Tests (pull_request) Successful in 3m52s
CI / Tests (push) Successful in 4m5s
build_models/build_critics parsed model_config into frozen dataclasses (ConditioningConfig, Stage2ModelConfig, ...) but then threw the parsed sub-objects away and passed the original raw dicts (conditioning["particle"], s2_spec.particle_type.to_dict()) down into ConditionEncoder/StageModel/etc, which re-read them with their own hardcoded .get(key, default) fallbacks — each an independent copy of a fact the dataclass already stated once. Worst instance: giant/training/trainers.py:236 converted an already-parsed ParticleTypeConfig back into a dict for no reason. Threads ConditioningAxisConfig (particle_cfg/material_cfg) and ParticleTypeConfig (particle_type_cfg) as the actual dataclass instances through every signature that used to type them dict: ConditionEncoder, StageModel/CriticModel, resolve_type_n_classes/stage2_type_dim/ stage2_trunk_sec_dim, giant/model/builders.py, giant/sample.py, giant/training/stage2_inputs.py, giant/training/trainers.py (StageSpec/ StageTrainer), giant/pipeline.py, giant/rollout.py, giant/validate.py — so ty now catches a misspelled field instead of it silently falling back. No config-schema change: config.toml/checkpoint model_config keep the same nested-dict shape; only what happens after the existing X.from_dict(...) parse changes. User-confirmed scope decision: both axes (particle_cfg/material_cfg and particle_type_cfg), not just the more heavily-duplicated particle_type_cfg axis, and not stopping at the two most literal parse-then-discard round trips — matching the issue's own proposal. Preserved-default decision: StageModel's particle_type_cfg=None sentinel (hit only by direct/test construction — build_models always passes an explicit particle_type) still resolves to ParticleTypeConfig(target= "physical"), not ParticleTypeConfig()'s own target="onehot" config-file default — switching it would have silently grown an unused, gradient-less type_head on every test that constructs Stage2OneShot/Stage2Autoregressive without particle_type_cfg=, breaking their "every param has a grad" checks. New tests in tests/test_network.py: ConditionEncoder/StageModel store the exact ConditioningAxisConfig/ParticleTypeConfig instance passed in (identity, not just equality) — no internal dict round-trip — and build_models's output carries real dataclass instances end to end, not the plain dicts it produced before this fix. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
74 lines
2.2 KiB
Python
74 lines
2.2 KiB
Python
import torch
|
|
from giant.config import ConditioningAxisConfig
|
|
from giant.constants import COND_DIM
|
|
from giant.model.network import Stage1Model
|
|
from giant.model.schedule import CosineSchedule, flow_matching_loss
|
|
from giant.sample import sample_flow, sample_ddim
|
|
|
|
PARTICLE_CFG = ConditioningAxisConfig(type="physical", emb_dim=8, n_layers=1)
|
|
MATERIAL_CFG = ConditioningAxisConfig(type="physical", emb_dim=8, n_layers=1)
|
|
|
|
|
|
def _small_model():
|
|
return Stage1Model(
|
|
pdg_vocab=3,
|
|
mat_vocab=2,
|
|
particle_cfg=PARTICLE_CFG,
|
|
material_cfg=MATERIAL_CFG,
|
|
hidden_dim=32,
|
|
n_res_blocks=2,
|
|
n_sec_head_k_max=15,
|
|
)
|
|
|
|
|
|
def _batch(B=8):
|
|
x1 = torch.randn(B, 9)
|
|
cond_cont = torch.randn(B, COND_DIM)
|
|
cond_cat = torch.zeros(B, 2, dtype=torch.long)
|
|
return x1, cond_cont, cond_cat
|
|
|
|
|
|
def test_flow_matching_loss_nonneg():
|
|
x1, cond_cont, cond_cat = _batch()
|
|
loss = flow_matching_loss(_small_model(), x1, cond_cont, cond_cat)
|
|
assert loss.item() >= 0.0
|
|
|
|
|
|
def test_flow_matching_loss_is_scalar():
|
|
x1, cond_cont, cond_cat = _batch()
|
|
loss = flow_matching_loss(_small_model(), x1, cond_cont, cond_cat)
|
|
assert loss.shape == ()
|
|
|
|
|
|
def test_flow_matching_loss_has_grad():
|
|
model = _small_model()
|
|
x1, cond_cont, cond_cat = _batch()
|
|
flow_matching_loss(model, x1, cond_cont, cond_cat).backward()
|
|
assert any(p.grad is not None for p in model.parameters())
|
|
|
|
|
|
def test_sample_flow_shape():
|
|
B = 6
|
|
cond_cont = torch.randn(B, COND_DIM)
|
|
cond_cat = torch.zeros(B, 2, dtype=torch.long)
|
|
sample, n_sec = sample_flow(_small_model(), cond_cont, cond_cat, steps=5)
|
|
assert sample.shape == (B, 9)
|
|
assert n_sec is not None and n_sec.shape == (B,)
|
|
|
|
|
|
def test_ddpm_loss_nonneg():
|
|
schedule = CosineSchedule(T=50)
|
|
x1, cond_cont, cond_cat = _batch()
|
|
loss = schedule.loss(_small_model(), x1, cond_cont, cond_cat)
|
|
assert loss.item() >= 0.0
|
|
|
|
|
|
def test_sample_ddim_shape():
|
|
B = 4
|
|
schedule = CosineSchedule(T=50)
|
|
cond_cont = torch.randn(B, COND_DIM)
|
|
cond_cat = torch.zeros(B, 2, dtype=torch.long)
|
|
sample, n_sec = sample_ddim(_small_model(), cond_cont, cond_cat, schedule, steps=5)
|
|
assert sample.shape == (B, 9)
|
|
assert n_sec is not None and n_sec.shape == (B,)
|