Maykeye

Maykeye

May 16

Baka4LLM. New horizons

#machinelearning #llm

1 min read

Maykeye

Apr 25

MambaBit. The most cursed LLM?

#machinelearning #llm

1 min read

Maykeye

Mar 30

BakaLLM 17.7 Simiplification!

1 min read

Maykeye

Mar 10

BakaLLM 17.6 More epochs! More evaluations! Cerebras slashed and smashed

1 min read

Maykeye

Feb 28

BakaLLM Part 17.5. Sssnake faster! Yet another milestone achieved

4 min read

Maykeye

Feb 21

BakaLLM, part 17. Dancing with SSMnakes

#machinelearning #ai #llm

2 min read

Maykeye

Feb 11

BakaLLM part NaN(still), Mamba shows promise and slowness

1 min read

Maykeye

Feb 4

BakaLLM part NaN and general shenanigans

1 min read

Maykeye

Jan 28

Part 16.1(x2) Copying RMT

1 min read

Maykeye

Jan 20

BakaLLM. Part 16. Procrastinating in style

#llm #machinelearning #ai

1 min read

Maykeye

Jan 10

BakaLLM, part 15.(inf). Why I can't hold all these stars or Pythia-70M is on the horizon

2 min read

Maykeye

Jan 6

BakaLLM, part 15.x: reachrf the stars, cooking on a plasma!

1 min read

Maykeye

Dec 29 '23

BakaLLM, part 15: reaching for the stars!

#machinelearning #ai #llm

2 min read

Maykeye

Dec 16 '23

BakaLLM, part 14: (MLP) Sharing is caring(?)

#machinelearning #llm #ai

2 min read

Maykeye

Dec 13 '23

BakaLLM step 13.1, achievement unlucked: flash attention

1 min read

Maykeye

Dec 9 '23

BakaLLM, step 13, RMT mainlined

1 min read

Maykeye

Dec 1 '23

BakaLLM, part 12.4, gatekeeping is all you need

#llm #machinelearning

3 min read

Maykeye

Nov 23 '23

BakaLLM, part 12, 1 step backward, 4 steps forward: starting new experiment

#llm #machinelearning

2 min read

Maykeye

Nov 19 '23

BakaLLM part 11.5. Moving away from (solely) WANDB

2 min read

Maykeye

Nov 16 '23

BakaLLM, part 11. Progress was made, this is not a DRILL, this is a PROGRESS! RMT and getting the loss <4.00, finally!

2 min read

Maykeye

Nov 9 '23

BakaLLM. Part 10. Progress? `Dim` light in thinning tunnel? Thin layers revisted: turning thin layers into thicc MLP

#llm #machinelearning

3 min read

Maykeye

Nov 5 '23

BakaLLM, part (9). The epoch of Stalling has begun!

#llm

4 min read

Maykeye

Oct 31 '23

BakaLLM, part 8. Let's <pause> for a second

#llm #ai #machinelearning

2 min read

Maykeye

Oct 26 '23

BakaLLM, part 7: XL struggles are done, for now

#llm #machinelearning

2 min read

Maykeye

Oct 18 '23

BakaLLM Part 6: Training and eXtra Large Struggles (Part 1) (hopefully last part of the struggles)

#llm #machinelearning

1

6 min read

Maykeye

Oct 13 '23

BakaLLM, p5: one stomp forward

#llm #machinelearning

3 min read

Maykeye

Oct 11 '23

BakaLLM, p4. Building LLM from scratch: New baseline is ready.

#llm #machinelearning

4 min read

Maykeye

Oct 6 '23

BakaLLM, part 3: it's testing time, it's testing time

#llm #machinelearning

1

4 min read

Maykeye

Sep 29 '23

BakaLLM, part 2

#llm #machinelearning

5 min read

Maykeye

Sep 25 '23

BakaLLM: Designing new LLM

#llm #machinelearning

6 min read

DEV Community

Badges

One Year Club

Writing Debut

8 Week Writing Streak

4 Week Writing Streak

Currently learning

Currently hacking on

Baka4LLM. New horizons

MambaBit. The most cursed LLM?

BakaLLM 17.7 Simiplification!

BakaLLM 17.6 More epochs! More evaluations! Cerebras slashed and smashed

BakaLLM Part 17.5. Sssnake faster! Yet another milestone achieved

BakaLLM, part 17. Dancing with SSMnakes

BakaLLM part NaN(still), Mamba shows promise and slowness

BakaLLM part NaN and general shenanigans

Part 16.1(x2) Copying RMT

BakaLLM. Part 16. Procrastinating in style

BakaLLM, part 15.(inf). Why I can't hold all these stars or Pythia-70M is on the horizon

BakaLLM, part 15.x: reachrf the stars, cooking on a plasma!

BakaLLM, part 15: reaching for the stars!

BakaLLM, part 14: (MLP) Sharing is caring(?)

BakaLLM step 13.1, achievement unlucked: flash attention

BakaLLM, step 13, RMT mainlined

BakaLLM, part 12.4, gatekeeping is all you need

BakaLLM, part 12, 1 step backward, 4 steps forward: starting new experiment

BakaLLM part 11.5. Moving away from (solely) WANDB

BakaLLM, part 11. Progress was made, this is not a DRILL, this is a PROGRESS! RMT and getting the loss <4.00, finally!

BakaLLM. Part 10. Progress? `Dim` light in thinning tunnel? Thin layers revisted: turning thin layers into thicc MLP

BakaLLM, part (9). The epoch of Stalling has begun!

BakaLLM, part 8. Let's <pause> for a second

BakaLLM, part 7: XL struggles are done, for now

BakaLLM Part 6: Training and eXtra Large Struggles (Part 1) (hopefully last part of the struggles)

BakaLLM, p5: one stomp forward

BakaLLM, p4. Building LLM from scratch: New baseline is ready.

BakaLLM, part 3: it's testing time, it's testing time

BakaLLM, part 2

BakaLLM: Designing new LLM