Akihito Miyazaki's picture

5 37

Akihito Miyazaki PRO

Akjava

·

https://github.com/akjava/

akjava

AI & ML interests

I’m creating a 3D/Voice Conversational AI Presentation Avatar with Diagrams

Recent Activity

liked a Space about 8 hours ago

ovi054/FLUX.Dev-LORA-Serverless

liked a Space 7 days ago

FunAudioLLM/ThinkSound

liked a Space 18 days ago

ovi054/image-to-prompt

View all activity

Organizations

None yet

Akjava 's Spaces 32

CreateConsistentCharacterFacialAnimationWithImg2Img

doc of img2img good for creating consistent animation

Mistral 7B Instruct v0.3 Matcha-TTS English

Enjoy TTS Chat

Matcha TTS Japanese

Description of Matcha TTS Japanese

Flux1 Schnell Img2img

Transform images based on text prompts

Flux.1 Schnell Mask Inpaint

Create or modify images by inpainting

Mediapipe Face Landmark/Skin Transform

Transform face landmark/skin,half of FaceSwap

Qwen2.5-0.5B-Rag-Thinking-llamacpp-cpu-rag-Flan-T5

Rag example with Qwen2.5-0.5B-Rag-Thinking +

Llamacpp-madlad400-3b-mt

madlad400-3b-mt on Llama.cpp CPU

Gemma3-4B-llamacpp-cpu-rag-smolagents

Rag example with Gemma3-4B and Smolagents

llhamacpp-flan-t5-large-grammar-synthesis

flan-t5-large-grammar-synthesis on Llama.cpp CPU

Smolagents ExtraSearchTools

Extra Search tool(google-custom,brave,priority) examples

chat-phi-4-deepseek-R1K-RL-EZO

chat with phi-4-deepseek-R1K-RL-EZO with CPU

Linear App Api Smolagents

smolagents example with linea.app api

Linear App Webhook Smolagents

Example of Direct Webhook-triggered AI Agent

Open Deep-Research-DuckDuckGo-Groq

OpenAI's Deep Research, but open with Groq(rate-limited)

Open Deep-Research-DuckDuckGo

OpenAI's Deep Research, but open without api key

Mediapipe Head Pose Estimation

2 head pose estimation with mediapipe and trained-model

Mediapipe Face Mesh 3d

create 3d-gltf face-mesh from image with mediapipe

Mediapipe Head 2d Spinning

spining mediapipe 468 points

Simple Go Emotions View Data

quick check for each emotion data

Mediapipe Change Eyes Direction

change eyes direction for guide-image

Mediapipe 68 Guide-Image Eyes-Closed and Mouth-Opened

create closed-eye and mouth opened guide image

Mediapipe 68 Points Facial Mask

create facial masks from 68 points landmark

Mediapipe 68 Points Facial Landmark

extract 68 points landmark from mediapipe-468