Zurück zu Fähigkeiten

analyze-generative-diffusion-model

pjt222
Aktualisiert 2 days ago
1 Ansichten
17
2
17
Auf GitHub ansehen
Metaaidesign

Über

Diese Fähigkeit analysiert vortrainierte generative Diffusionsmodelle wie Stable Diffusion, indem Qualitätsmetriken (FID, CLIP-Score) berechnet, Aufmerksamkeitskarten visualisiert und latente Räume untersucht werden. Nutzen Sie sie, um die Ausgabequalität von Modellen zu bewerten, Rauschzeitpläne zu vergleichen oder Kreuzaufmerksamkeitsmuster für textkonditionierte Generierung zu analysieren. Sie ist für Entwickler konzipiert, die erweiterte Modellbewertung und -inspektion durchführen.

Schnellinstallation

Claude Code

Empfohlen
Primär
npx skills add pjt222/agent-almanac -a claude-code
Plugin-BefehlAlternativ
/plugin add https://github.com/pjt222/agent-almanac
Git CloneAlternativ
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/analyze-generative-diffusion-model

Kopieren Sie diesen Befehl und fügen Sie ihn in Claude Code ein, um diese Fähigkeit zu installieren

Dokumentation

Analyze a Generative Diffusion Model

Evaluate pre-trained generative diffusion models. Quantitative quality metrics, noise schedule inspection, cross-attention map analysis, latent space probing. Understand model behavior, diagnose failure modes, guide fine-tuning decisions.

When Use

  • Evaluating pre-trained generative diffusion model's output quality with standard metrics
  • Computing FID, IS, CLIP score, or precision/recall for generated image sets
  • Inspecting and comparing noise schedules (linear, cosine, learned) via SNR curves
  • Extracting cross-attention maps to understand text-to-image token-region correspondences
  • Interpolating between latent codes or discovering semantic directions in latent space
  • Detecting out-of-distribution inputs for diffusion model pipeline

Inputs

  • Required: Pre-trained model identifier or checkpoint path (e.g., stabilityai/stable-diffusion-2-1)
  • Required: Analysis mode — one or more of: metrics, schedule, attention, latent
  • Required: Reference dataset for metric computation (real images or dataset name)
  • Optional: Text prompts for attention analysis (default: model-appropriate test prompts)
  • Optional: Number of generated samples for metric computation (default: 10000)
  • Optional: Device configuration (default: cuda if available, else cpu)

Steps

Step 1: Quantitative Evaluation

Compute standard generative quality metrics against reference dataset.

  1. Set up evaluation pipeline:
import torch
from diffusers import StableDiffusionPipeline
from torchmetrics.image.fid import FrechetInceptionDistance
from torchmetrics.image.inception import InceptionScore

device = "cuda" if torch.cuda.is_available() else "cpu"
pipe = StableDiffusionPipeline.from_pretrained(
    "stabilityai/stable-diffusion-2-1", torch_dtype=torch.float16
).to(device)

fid = FrechetInceptionDistance(feature=2048, normalize=True).to(device)
inception = InceptionScore(normalize=True).to(device)
  1. Feed real images into metric accumulators:
from torch.utils.data import DataLoader

for batch in DataLoader(real_dataset, batch_size=64):
    imgs = (batch * 255).byte().to(device)
    fid.update(imgs, real=True)
  1. Generate samples and accumulate fake statistics:
prompts = load_evaluation_prompts("prompts.txt")  # one prompt per line
n_generated = 0
while n_generated < 10000:
    prompt_batch = prompts[n_generated:n_generated + 8]
    images = pipe(prompt_batch, num_inference_steps=50).images
    tensors = torch.stack([to_tensor(img) for img in images]).to(device)
    byte_imgs = (tensors * 255).byte()
    fid.update(byte_imgs, real=False)
    inception.update(byte_imgs)
    n_generated += len(images)
  1. Compute CLIP score for text-image alignment:
from torchmetrics.multimodal.clip_score import CLIPScore

clip_metric = CLIPScore(model_name_or_path="openai/clip-vit-large-patch14").to(device)
for prompt, image_tensor in zip(sampled_prompts, sampled_tensors):
    clip_metric.update(image_tensor.unsqueeze(0), [prompt])

print(f"FID: {fid.compute():.2f}")
print(f"IS:  {inception.compute()[0]:.2f} +/- {inception.compute()[1]:.2f}")
print(f"CLIP: {clip_metric.compute():.2f}")
  1. Compute precision and recall for mode coverage:
from torchmetrics.image import FrechetInceptionDistance

# Precision: fraction of generated images near real manifold
# Recall: fraction of real images near generated manifold
# Use improved precision/recall (Kynkaanniemi et al., 2019) via
# feature embeddings from the Inception network

Got: FID below 30 for well-trained Stable Diffusion model on standard benchmarks. IS above 50 on ImageNet-class prompts. CLIP score above 25 for text-conditioned models. Precision and recall both above 0.6.

If fail: FID above 100? Verify real and generated images share same resolution and normalization. CLIP score low but FID acceptable? Model generates plausible images that do not match text prompt -- check text encoder. Ensure at least 10,000 samples for stable FID estimates.

Step 2: Noise Schedule Inspection

Visualize and compare forward and reverse noise schedules.

  1. Extract schedule parameters from model:
scheduler = pipe.scheduler
betas = torch.tensor(scheduler.betas) if hasattr(scheduler, 'betas') else None
alphas_cumprod = torch.tensor(scheduler.alphas_cumprod)
timesteps = torch.arange(len(alphas_cumprod))
  1. Compute signal-to-noise ratio curve:
import numpy as np
import matplotlib.pyplot as plt

snr = alphas_cumprod / (1 - alphas_cumprod)
log_snr = torch.log(snr)

fig, axes = plt.subplots(1, 3, figsize=(18, 5))
axes[0].plot(timesteps.numpy(), alphas_cumprod.numpy())
axes[0].set_xlabel("Timestep"); axes[0].set_ylabel("alpha_cumprod")
axes[0].set_title("Cumulative Signal Retention")

axes[1].plot(timesteps.numpy(), log_snr.numpy())
axes[1].set_xlabel("Timestep"); axes[1].set_ylabel("log(SNR)")
axes[1].set_title("Log Signal-to-Noise Ratio")

if betas is not None:
    axes[2].plot(timesteps.numpy(), betas.numpy())
    axes[2].set_xlabel("Timestep"); axes[2].set_ylabel("beta")
    axes[2].set_title("Beta Schedule")
fig.tight_layout()
fig.savefig("noise_schedule.png", dpi=150)
  1. Compare multiple schedule types:
from diffusers import DDPMScheduler

schedules = {
    "linear": DDPMScheduler(beta_schedule="linear", num_train_timesteps=1000),
    "cosine": DDPMScheduler(beta_schedule="squaredcos_cap_v2", num_train_timesteps=1000),
}

fig, ax = plt.subplots(figsize=(10, 6))
for name, sched in schedules.items():
    ac = torch.tensor(sched.alphas_cumprod)
    snr = torch.log(ac / (1 - ac))
    ax.plot(snr.numpy(), label=name)
ax.set_xlabel("Timestep"); ax.set_ylabel("log(SNR)")
ax.set_title("Schedule Comparison"); ax.legend()
fig.savefig("schedule_comparison.png", dpi=150)

Got: Cosine schedule shows more gradual SNR decrease in mid-timesteps compared to linear. Log-SNR curve should span from approximately +10 (clean) to -10 (pure noise). Learned schedules should be monotonically decreasing.

If fail: alphas_cumprod not monotonically decreasing? Schedule misconfigured. Values constant? Check scheduler properly initialized with model's config. For custom schedulers, verify set_timesteps() called.

Step 3: Attention Map Analysis

Extract and visualize cross-attention maps from text-conditioned models.

  1. Register attention hooks on U-Net cross-attention layers:
attention_maps = {}

def hook_fn(name):
    def fn(module, input, output):
        # Cross-attention: Q from image, K/V from text
        if hasattr(module, 'processor'):
            attention_maps[name] = output.detach().cpu()
    return fn

for name, module in pipe.unet.named_modules():
    if 'attn2' in name and hasattr(module, 'processor'):
        module.register_forward_hook(hook_fn(name))
  1. Run inference and collect attention at specific timesteps:
prompt = "a red car parked next to a blue house"
timestep_attention = {}

# Custom callback to capture attention at specific timesteps
def callback_fn(pipe, step_index, timestep, callback_kwargs):
    if step_index in [5, 15, 30, 45]:
        timestep_attention[int(timestep)] = {
            k: v.clone() for k, v in attention_maps.items()
        }
    return callback_kwargs

output = pipe(prompt, num_inference_steps=50, callback_on_step_end=callback_fn)
  1. Visualize token-region correspondences:
tokenizer = pipe.tokenizer
tokens = tokenizer.encode(prompt)
token_strings = [tokenizer.decode([t]) for t in tokens]

# Select a mid-resolution attention layer
layer_key = [k for k in attention_maps if 'mid' in k or 'up.1' in k][0]
attn = attention_maps[layer_key]  # shape: (batch, heads, hw, seq_len)
attn_avg = attn.mean(dim=1)  # average across heads
res = int(attn_avg.shape[1] ** 0.5)
attn_map = attn_avg[0].reshape(res, res, -1)

fig, axes = plt.subplots(2, min(len(token_strings), 6), figsize=(18, 6))
for idx, token in enumerate(token_strings[:6]):
    for row, (ts, ts_attn) in enumerate(list(timestep_attention.items())[:2]):
        a = ts_attn[layer_key].mean(dim=1)[0]
        a_res = int(a.shape[0] ** 0.5)
        axes[row, idx].imshow(a[:, idx].reshape(a_res, a_res), cmap="hot")
        axes[row, idx].set_title(f"t={ts}: '{token}'")
        axes[row, idx].axis("off")
fig.suptitle("Cross-Attention Maps by Token and Timestep")
fig.tight_layout()
fig.savefig("attention_maps.png", dpi=150)

Got: Content tokens ("car", "house") activate localized spatial regions. Style/color tokens ("red", "blue") activate regions overlapping with their associated object. Early timesteps (high noise) show diffuse attention; later timesteps show sharp, localized attention.

If fail: All attention maps look uniform? Hook may be capturing self-attention instead of cross-attention -- verify layer name contains attn2 (cross) not attn1 (self). Attention captured but has wrong dimensions? Check output tensor indexing matches layer's head count and spatial resolution.

Step 4: Latent Space Probing

Explore structure of latent space through interpolation and direction discovery.

  1. Encode reference images into latent space:
from diffusers import AutoencoderKL
from PIL import Image
import torchvision.transforms as T

vae = pipe.vae
transform = T.Compose([T.Resize(512), T.CenterCrop(512), T.ToTensor(),
                       T.Normalize([0.5], [0.5])])

def encode_image(image_path):
    img = transform(Image.open(image_path).convert("RGB")).unsqueeze(0).to(device)
    with torch.no_grad():
        latent = vae.encode(img.half()).latent_dist.sample() * vae.config.scaling_factor
    return latent

z1 = encode_image("image_a.png")
z2 = encode_image("image_b.png")
  1. Perform spherical linear interpolation (slerp):
def slerp(z1, z2, alpha):
    """Spherical linear interpolation between two latent codes."""
    z1_flat = z1.flatten()
    z2_flat = z2.flatten()
    omega = torch.acos(torch.clamp(
        torch.dot(z1_flat, z2_flat) / (z1_flat.norm() * z2_flat.norm()), -1, 1
    ))
    if omega.abs() < 1e-6:
        return (1 - alpha) * z1 + alpha * z2
    return (torch.sin((1 - alpha) * omega) * z1 + torch.sin(alpha * omega) * z2) / torch.sin(omega)

alphas = torch.linspace(0, 1, 8)
interpolated = [slerp(z1, z2, a.item()) for a in alphas]
decoded = []
for z in interpolated:
    with torch.no_grad():
        img = vae.decode(z / vae.config.scaling_factor).sample
    decoded.append(img.cpu())
  1. Discover semantic directions via prompt-pair differences:
def get_text_embedding(prompt):
    tokens = pipe.tokenizer(prompt, return_tensors="pt", padding="max_length",
                            max_length=77, truncation=True).input_ids.to(device)
    with torch.no_grad():
        emb = pipe.text_encoder(tokens).last_hidden_state
    return emb

pos_emb = get_text_embedding("a happy person smiling")
neg_emb = get_text_embedding("a sad person frowning")
direction = pos_emb - neg_emb  # semantic direction in text embedding space
  1. Detect out-of-distribution latents:
# Compute latent space statistics from a reference set
ref_latents = torch.stack([encode_image(p) for p in reference_paths])
ref_mean = ref_latents.mean(dim=0)
ref_std = ref_latents.std(dim=0)

def ood_score(z):
    """Mahalanobis-like OOD score (higher = more unusual)."""
    deviation = ((z - ref_mean) / (ref_std + 1e-6)).flatten()
    return deviation.norm().item()

test_z = encode_image("test_image.png")
score = ood_score(test_z)
print(f"OOD score: {score:.2f} (reference mean: {np.mean([ood_score(r) for r in ref_latents]):.2f})")

Got: Interpolated images show smooth, semantically meaningful transitions without artifacts. Semantic directions produce consistent attribute changes when added to diverse latent codes. OOD scores for in-distribution images cluster tightly; outliers score significantly higher.

If fail: Interpolation produces blurry or incoherent midpoints? Use slerp instead of linear interpolation -- linear interpolation traverses low-density regions in high-dimensional latent spaces. Semantic directions have no visible effect? Increase direction magnitude or verify text encoder is same one used during model training.

Checks

  • FID computed on at least 10,000 generated samples and matching real sample count
  • CLIP score computed with same CLIP model used during training (if applicable)
  • Noise schedule visualization shows monotonically decreasing alphas_cumprod
  • Log-SNR spans approximately +10 to -10 across full timestep range
  • Attention maps resolve per-token spatial activations at mid-resolution layers
  • Attention sharpens from early (diffuse) to late (localized) timesteps
  • Latent interpolations smooth with no sudden jumps or artifacts
  • OOD detection baseline established from at least 100 reference samples

Pitfalls

  • FID on mismatched resolutions: Real and generated images must be same resolution before feeding to Inception. Resize both sets identically or FID will be inflated.
  • Forgetting to normalize for torchmetrics: FrechetInceptionDistance(normalize=True) expects [0, 1] float tensors. With normalize=False it expects [0, 255] uint8. Mixing conventions gives meaningless FID.
  • Hooking self-attention instead of cross-attention: U-Net layers named attn1 are self-attention (image-to-image). Use attn2 for cross-attention (text-to-image). Confusing them produces uninformative uniform maps.
  • Linear interpolation in high dimensions: Linear interpolation between two high-dimensional Gaussians passes through low-density shell. Always use slerp for latent space interpolation in diffusion models.
  • Ignoring VAE scaling factor: Stable Diffusion latents scaled by vae.config.scaling_factor after encoding. Forgetting to apply or remove this factor produces garbled decoded images.
  • Too few samples for precision/recall: Precision and recall estimates from fewer than 5,000 samples per set unreliable. Use at least 10,000 for stable estimates.

See Also

  • implement-diffusion-network - building diffusion models that this skill evaluates
  • analyze-diffusion-dynamics - mathematical foundations of noise processes inspected here
  • fit-drift-diffusion-model - different diffusion model family sharing SDE foundations

GitHub Repository

pjt222/agent-almanac
Pfad: i18n/caveman/skills/analyze-generative-diffusion-model
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

Verwandte Skills

content-collections

Meta

Diese Skill bietet eine produktionsgetestete Einrichtung für Content Collections – ein TypeScript-first-Tool, das Markdown/MDX-Dateien in typsichere Datensammlungen mit Zod-Validierung umwandelt. Verwenden Sie ihn beim Erstellen von Blogs, Dokumentationsseiten oder inhaltsstarken Vite + React-Anwendungen, um Typsicherheit und automatische Inhaltsvalidierung zu gewährleisten. Er behandelt alles von der Vite-Plugin-Konfiguration und MDX-Kompilierung bis hin zur Deployment-Optimierung und Schema-Validierung.

Skill ansehen

polymarket

Meta

Diese Fähigkeit ermöglicht es Entwicklern, Anwendungen mit der Polymarket-Prognosemärkte-Plattform zu erstellen, einschließlich API-Integration für Handel und Marktdaten. Sie bietet außerdem Echtzeit-Datenstreaming über WebSocket, um Live-Trades und Marktaktivitäten zu überwachen. Nutzen Sie sie zur Implementierung von Handelsstrategien oder zur Erstellung von Tools, die Live-Marktaktualisierungen verarbeiten.

Skill ansehen

creating-opencode-plugins

Meta

Diese Fähigkeit unterstützt Entwickler dabei, OpenCode-Plugins zu erstellen, die in über 25 Ereignistypen wie Befehle, Dateien und LSP-Operationen eingreifen. Sie bietet die Plugin-Struktur, Event-API-Spezifikationen und Implementierungsmuster für JavaScript/TypeScript-Module. Nutzen Sie sie, wenn Sie den Lebenszyklus des OpenCode KI-Assistenten mit benutzerdefinierter ereignisgesteuerter Logik abfangen, überwachen oder erweitern müssen.

Skill ansehen

sglang

Meta

SGLang ist ein hochperformantes LLM-Serving-Framework, das sich auf schnelle, strukturierte Generierung für JSON, Regex und agentenbasierte Workflows unter Verwendung seines RadixAttention-Prefix-Cachings spezialisiert. Es bietet deutlich schnellere Inferenz, insbesondere für Aufgaben mit wiederholten Präfixen, was es ideal für komplexe, strukturierte Ausgaben und Mehrfachdialoge macht. Wählen Sie SGLang gegenüber Alternativen wie vLLM, wenn Sie constrained decoding benötigen oder Anwendungen mit umfangreicher Präfix-Weitergabe entwickeln.

Skill ansehen