We got amazing new discount coupon from Massed Compute - SECourses coupon to get half price for L40 GPUs - 48 GB Just Amazing
Added 2024-12-17 01:32:24 +0000 UTCAs you know Massed Compute is the very best Cloud Service provider at the moment with the fastest Disk speeds and Network speeds.
They just gave us a new coupon to use on L40 GPUs.
Now you can use both RTX A6000, RTX A6000 ALT and L40 GPUs.
Please register for Massed Compute using the following link:
https://vm.massedcompute.com/signup?linkId=lp_034338&sourceId=secourses&tenantId=massed-compute
Use our coupon : SECourses
L40 from 0.99 $ to 0.49 $ per hour
A6000 from 0.62 $ to 0.31 $ per hour
A6000 ALT Config (only has 32 GB RAM instead of 48 GB RAM) : from 0.6 $ to 0.3 $ per hour
Comments
You are welcome and yes :)
Furkan Gözükara
2024-12-26 20:24:50 +0000 UTCThank you so much and welcome. Currently I see opportunities in learning how to train something consistent. It can be art style, object, subject and such.
Furkan Gözükara
2024-12-26 20:24:42 +0000 UTCthank you for the new L40 Coupon, amazing speed
Shepard4k
2024-12-26 16:08:00 +0000 UTCThank you for teaching us all these amazing methods and making it more accessible. =) I love your accent and I understand every word of it. It took me a long time to finally become a member and it would have been easier and cheaper in the long run if I had. So many guides out there it gets too much with various different opinions. I haven't seen anyone that worked so hard and systematically as you have. Do you have any advice on any AI tool I should learn in order to get a new job? With that said, I'm far away from your level of expertise and knowledge and have no illusions that I will ever become as skilled as you are. Take care. /Hockey from Sweden
Hockey
2024-12-26 15:55:34 +0000 UTCyou are welcome. we also upgraded image and now faster. some scripts may need update hopefully i will do over time
Furkan Gözükara
2024-12-22 22:24:29 +0000 UTCThank you! Amazing indeed. I switched to using Massed Compute instead of dealing with my windows pc and rtx2080 errors. Now it works wonders with your packages. Thank you for your service
Mehmet Atakan Çavuşlu
2024-12-22 21:11:01 +0000 UTCwill make installers hopefully thanks
Furkan Gözükara
2024-12-18 08:29:46 +0000 UTCthanks you, by the way have you seen this new model? it looks really good https://github.com/lehduong/OneDiffusion
Ryokazuma
2024-12-18 07:34:27 +0000 UTCawesome
Furkan Gözükara
2024-12-17 21:38:12 +0000 UTCfloat will make it 2x size but bf16 and fp16 same size
Furkan Gözükara
2024-12-17 21:38:04 +0000 UTCI'm trying it right now, it's a little bit faster 5,80 it/s with your best quality config (48GB_GPU_28200MB_6.3_second_it_Tier_1) with more hard drive space for the price it's amazing, Thanks again!!!! I have another question if i can, on my computer i have an rtx 2080 super it can run a file trained with your best config but it is really slow so i was trying to train an flux dev fp8 but the trained files is still 23.8gb, i was wondering if i train it in float or bf16 will it change the size to the same size as the fp8? Thank you
Ryokazuma
2024-12-17 21:18:31 +0000 UTCAre the L40 as fast as the A6000 for training?
Ryokazuma
2024-12-17 14:38:01 +0000 UTCThank you very much!!
Ryokazuma
2024-12-17 14:37:05 +0000 UTCThank you very much! I need to use this and try out some of the higher-quality Flux training configs. ε(´｡•᎑•`)っ 💕
reaper557
2024-12-17 14:33:33 +0000 UTCL40 is still plenty until it also dries :D
Furkan Gözükara
2024-12-17 12:14:10 +0000 UTCThat's goog... unfortunately it's very hard to find some GPU's available lately... but it's so good to get this stuf at these prices...
Art
2024-12-17 12:06:17 +0000 UTCYou are welcome
Furkan Gözükara
2024-12-17 11:58:20 +0000 UTCThank you Doc.
Rollin90sCrips
2024-12-17 10:35:04 +0000 UTC
