kimi000/crystal-field-29

FLUX.2 Klein 4B Step-400 EMA Diffusers Export

This repository contains a complete native Hugging Face Diffusers export of a trained FLUX.2 Klein Base 4B checkpoint. The public repository name is intentionally neutral; full experiment provenance is recorded below for reproducibility.

Provenance

• Source experiment: flux2_klein_base_4b_diffusionnft_dvreward_prompt_rubric_v4_3_1_geneval_16prompts_group14_7train_1dvreward_tp1_2node_512px_20step_cfg4_nrt
• Experiment variant: prompt-rubric v4.3.1, GenEval
• Formal run: flux2-v431-geneval-formal-505-142eef2e-a12
• Source run code revision: 142eef2e8ea30d30eaff17e3ddb2547f908fc3cc
• Source checkpoint: models/step_400.pt
• Global step: 400
• Exported weights: EMA
• Source checkpoint SHA-256: 7aca3ce4549f66635e82e6ee7e6bedad9ba9d1193defc16ca59aeb0457015ad1
• Base model: black-forest-labs/FLUX.2-klein-base-4B
• Base revision: a3b4f4849157f664bdbc776fd7453c2783562f4d
• Training profile: 16 prompts, group size 14, 512 px, 20 rollout steps, CFG 4, 7 policy ranks and 1 DVReward replica per node across 2 nodes
• LoRA rank / alpha: 32 / 64

The EMA LoRA weights were merged into the native BF16 FLUX.2 transformer. The export contains 169 transformer tensors and 3,875,544,576 transformer parameters. DiffusionNFT old-policy training-state tensors were excluded from the inference artifact.

Format

• Pipeline: diffusers.Flux2KleinPipeline
• Transformer: diffusers.Flux2Transformer2DModel
• Precision: BF16
• Scheduler: 1,000 training timesteps, dynamic shifting enabled, deterministic sampling
• Transformer weights: 9 shards
• Text encoder weights: 9 shards
• Maximum shard size: below 1 GB

Verification

The export was strictly reloaded offline with stock Diffusers and generated a 512 x 512 smoke image using 4 inference steps, CFG 4, and seed 0. The merged transformer differs from Base in 60 tensors and 2,403,252,754 elements, with L2 delta 26.236759527279407 and maximum absolute delta 0.007568359375.

A same-seed Base render was also generated:

• Changed channel values: 763,458
• Mean absolute channel delta: 39.71453730265299
• Maximum absolute channel delta: 255

These checks verify that the exported transformer contains the trained Step-400 EMA weights rather than an unchanged Base transformer. verification.json, smoke images, the export manifest, and SHA-256 checksums are included.

Loading

import torch
from diffusers import Flux2KleinPipeline

pipe = Flux2KleinPipeline.from_pretrained(
    "kimi000/crystal-field-29",
    torch_dtype=torch.bfloat16,
)
pipe.enable_model_cpu_offload()

image = pipe(
    prompt="A red cube beside a blue glass sphere.",
    height=1024,
    width=1024,
    guidance_scale=4.0,
    num_inference_steps=20,
).images[0]

See export_manifest.json and verification.json for machine-readable export details.