5 ms·
> We have been looking forward to reducing pricing for a long time. Our teams have made incredible progress in making our models more efficient to run, which ha
by AJRF 4y ago
> We have been looking forward to reducing pricing for a long time. Our teams have made incredible progress in making our models more efficient to run, which has reduced the cost it takes to serve them, and we are now passing these savings along to our customers.
Nothing to do with Stable Diffusion then, no?
- JimmyAustin 4y agoGiven that Dall-E 2's pricing hasn't changed, probably not.
- aaroninsf 4y agoNow that StableDiffusion is released, I expect that to change, unless there is a (certainly ultimately doomed) attempt to Hold the Line.
- cma 4y agoCould be the GPU price crash as well (even if they use Google's TPUs, TPU demand would presumably go down in a GPU price crash).
- tomComb 4y agoMicrosoft bought control of OpenAI so I believe this would all be running on Azure (so would not be using Google TPU's).
- tvhahn 4y agoWhy bother use a GPT-3 model when you can use the free BLOOM model? Large "foundation" models will become a commodity -- a race to the bottom.
- grej 4y agoI have used BLOOM a little, and I struggle to get it to produce the same kind of output as GPT-3. Anyone have tips there? It often seems to not keep the right context of the previously generated text.
- aljungberg 4y agoGPT-3 has been fine-tuned after release to better interpret prompts (see InstructGPT). Perhaps Bloom is more like the original GPT-3; a little more raw and requiring better prompt engineering? In my small amount of testing of Bloom so far it seems capable of advanced behaviour but it can indeed be trickier to coax that out. Playing with temperature and sampling matters for sure.
- juliensalinas 4y agoI've been testing BLOOM for a while and it seems it is working very well with good few-shot learning. See this post about prompt examples: https://nlpcloud.com/effectively-using-gpt-j-gpt-neo-gpt-3-alternatives-few-shot-learning.html https://nlpcloud.com/effectively-using-gpt-j-gpt-neo-gpt-3-a... All of these examples work great with BLOOM. But for the moment you can't just use pure natural language instructions with BLOOM indeed. Maybe it will come!
- choppaface 4y agoIt’s not just Stable Diffusion the model, but the research is become really actually open unlike “open ai”
- lee101 4y agoYes this move is I think forced hand by competitors like https://text-generator.io https://text-generator.io etc which is already 10x cheaper. If the open source community then other company's didn't follow along with them then we would still be being price gauged in a monopoly situation, hopefully same thing happens to dalle soon