Sign In

Buratino & Co [Hunyuan Video]

28
108
8
Updated: Jan 5, 2025
charactervideohunyuan
Type
LoRA
Stats
108
0
Reviews
Published
Jan 5, 2025
Base Model
Hunyuan Video
Training
Steps: 4,000
Epochs: 50
Trigger Words
Buratino
Malvina
Alisa
Bazilio
Hash
AutoV2
30AA7E301C
default creator card background decoration
GO
Gorro17
Tencent Hunyuan is licensed under the Tencent Hunyuan Community License Agreement, Copyright © 2024 Tencent. All Rights Reserved. The trademark rights of “Tencent Hunyuan” are owned by Tencent or its affiliate.
Powered by Tencent Hunyuan

This LoRA was created using frames from the Soviet film "The Adventures of Buratino" (1975), an adaptation of the Soviet version of Carlo Collodi's book "The Adventures of Pinocchio".

At first, I made a version featuring only one character - Buratino. However, I decided to test whether it would be possible to create a LoRA for multiple characters simultaneously. It turned out to be quite feasible! This LoRA can depict four characters from the film: the boy with the long nose - Buratino, the girl with blue hair named Malvina, the red-haired woman - Alisa, and the man with dark glasses and cat-like whiskers - Bazilio.

An important condition for proper functioning is providing an accurate description of the character, including their name and details.

Buratino:

A boy Buratino with curly blond hair and a long, pointed nose. He wears a red-and-white striped cap with a pom-pom at the end and a yellow shirt.

Malvina:

A girl Malvina with curly light-blue hair and a large white bow. She is dressed in a white dress.

Alisa:

A woman Alisa with red hair who wears a hat with feathers and a fur coat.

Bazilio:

A man Bazilio who wears a hat and round sunglasses. He has sideburns and cat-like whiskers.

I managed to generate a video featuring all four characters simultaneously. However, most of the time, when there are more than two characters, details tend to get mixed up. For instance, you might end up with a blue-haired girl sporting a long nose, sunglasses, and a feathered hat.

The LoRA was trained on horizontal frames, so black bars may appear at the top and bottom of vertical frames. Additionally, in some cases, increasing the LoRA weight to 1.2 can help achieve a closer resemblance of the characters to the original.

Of course, the film featured more characters, and if there is interest in this LoRA, I will try to create a version that includes all the characters from the movie.