F.1 Basic Algorithm Model
Back

F.1 Basic Algorithm Model
2267 23 297

Realistic

F.1 Basic Algorithm Model

Usage in WEBUI:

 The fp8 version is recommended, as it is faster and more stable.

The fp16 version is slightly slower in loading and more likely to cause out-of-memory errors.

schnell is the turbo version, with average performance, not recommended.


Currently, the sampling method in WEBUI only supports FlowMatchEuler. Embedding is compatible with SD1.5, but there may be issues with SDXL (not absolute).

The official recommendation for cfg is 3.5, but users can adjust it as needed.

The default sampling steps are 20, which can be increased as appropriate.


Usage in ComfyUI:

There are more combinational options, allowing for extended gameplay through stream-based pipelines.

The standard version requires a computer with at least 32GB of RAM. Testing shows that the 4090 fully utilizes its VRAM. The dev fp8 version is recommended for local use.

The outputs generated by this model can be used for personal, scientific, and research purposes, and must comply with the commercial licensing of the original model.

This model is sourced from an external transfer (transfer address: https://www.liblib.art/modelinfo/488cd9d58cd4421b9e8000373d7da123?from=pic_detail&versionUuid=412b427ddb674b4dbab9e5abd5ae6057 ）,if the original author has objections to this transfer, you can click，
Appeal
We will, within 24 hours, edit, delete, or transfer the model to the original author according to the original author's request

user_b0f988dd1f60489892e1b20899036b0c

user_b0f988dd1f60489892e1b20899036b0c

Realistic

Model Information

Active
Original author：
热点推荐官
Model Type：
Checkpoint
Basic Model：
F1基础 D
Resource Name：
models/checkpoints/F.1基础算法模型-F.1-dev-fp8.safetensors
MD5：
8cb01185797e62db53258164835f0d3a

Usage in WEBUI:

 The fp8 version is recommended, as it is faster and more stable.

The fp16 version is slightly slower in loading and more likely to cause out-of-memory errors.

schnell is the turbo version, with average performance, not recommended.


Currently, the sampling method in WEBUI only supports FlowMatchEuler. Embedding is compatible with SD1.5, but there may be issues with SDXL (not absolute).

The official recommendation for cfg is 3.5, but users can adjust it as needed.

The default sampling steps are 20, which can be increased as appropriate.


Usage in ComfyUI:

There are more combinational options, allowing for extended gameplay through stream-based pipelines.

The standard version requires a computer with at least 32GB of RAM. Testing shows that the 4090 fully utilizes its VRAM. The dev fp8 version is recommended for local use.

The outputs generated by this model can be used for personal, scientific, and research purposes, and must comply with the commercial licensing of the original model.

This model is sourced from an external transfer (transfer address: https://www.liblib.art/modelinfo/488cd9d58cd4421b9e8000373d7da123?from=pic_detail&versionUuid=412b427ddb674b4dbab9e5abd5ae6057 ）,if the original author has objections to this transfer, you can click，
Appeal
We will, within 24 hours, edit, delete, or transfer the model to the original author according to the original author's request