A good overview of LangChain by @Saboo_Shubham_ ! Thanks for the write up 🙂
TOOLS
-
ASR Corpus Creator Utility by Yehor Smoliakov
By
–
You can also use awesome utilities like the one @yehor_smoliakov has created: https://
github.com/egorsmkv/asr-c
orpus-creator
… -
Fine-tuning Speech Recognition Models with Open Datasets
By
–
This is a challenge for sure. However, in some cases it is possible to find an academic/ open source dataset for the language that you’re interested in fine tuning. There is also always the possibility of using a Wav2Vec2.0 checkpoint to transcribe YT videos & use that.
-
Open Source Speech Recognition Platform Democratizing Access
By
–
We are on a mission to democratise speech, increase the language coverage of current SoTA speech recognition and push the limits of what is possible. From beginners to hobbyists to academics to power users, come one, come all! Register here: https://
forms.gle/QAkYnnSFgGb84d
UX6
… -
Developer Upgrades E-Commerce Search Engine with Deep Learning
By
–
This is amazing – a single awesome dev in his spare time has massively upgraded a giant e-commerce company's search engine, thanks to deep learning and #nbdev https://
t.co/pANohhY7mI -
Comparing Stable Diffusion 1.5 vs 2.0 Image Generation Quality
By
–
easy to compare a lot of images from both models on http://
stableboost.ai , e.g. "cute dog cooking tacos, photorrealistic", grid of boosted images from 1.5 (left) and 2.0 (right). 2.0 looking more distorted, cartoony, simpler, ignores text more. may need more prompt engineering -
Stable Diffusion 2.0 Shows Quality Decline Compared to 1.5
By
–
plot twist: stable diffusion 2.0 looks quite a bit worse on the few prompts i've tried so far compared to 1.5 (even not including celebrities/artists). Running theory seems to be this is due to an aggressive data sanitization campaign since the original release (?).
-
Code Injection in Notebooks for Parameterization
By
–
Have you checked out the ability also to inject code into a cell to parameterise notebooks?
-
Current Limitations of Image Output Sizes and Scheduler Options
By
–
It doesn't yet support a few things. It only outputs 512 and 768 sized images, doesn't support init_image or mask, and it only has Euler and DDIM schedulers.
-
Stable Diffusion 2.0 Release Now Available on Replicate
By
–
Congrats @StabilityAI on the Stable Diffusion 2.0 release. We've got an early version to try on Replicate: