<p>Flux 1 Dev likeness of MaryLee, a fictional Asian American woman.</p><p><a rel="ugc" href="https://civarchive.com/models/32786/mary-lee">Stable Diffusion version</a></p><h3 id="notes-after-a-week-of-use.-cv3a1m0zx">Notes after a week of use.</h3><p>This is my first LoRA for Flux, and it is highly dependent on the same training I used for the same likeness for SD. As such, it does lean a bit into the plastics type look, especially at higher strengths.</p><p>This is also a relatively "weak" likeness model. Compared to my earlier versions in SD, it's a poor direct and consistent training, and acts more like a strong suggestion than a clone.</p><h2 id="one-weird-thing-about-flux-this-lora-is-good-for.-uwksn5gro">One weird thing about Flux this LoRA is good for.</h2><p>FLUX 1 has an interesting habit of having maybe 100 or so likenesses in the base training. You see this over time, with the same few faces (shapes, features, expressions) being the default for a gender + ethnicity + age.</p><p><strong>MaryLee v1 is luckily pretty similar to one of these base likeness, and seems to be consistent in triggering that same vague face.</strong></p>

MaryLee

<p>Updated training, with expanded references (80 vs. 20 in v1.0) and GPT-4o tagging based on updated community notes around Flux's tagging system.</p>

Files with SHA256: 3c5e30c7fb5f7364330ac65c2f150a98a8d302c6f9b5333230a64094c8c7ce75