rasbt
713a6e24c9
add tests
2025-06-22 17:48:23 -05:00
martinzwm
ffc5e4e5d6
Fix bug in masking when kv cache is used.
2025-06-22 14:06:00 -07:00
Sebastian Raschka
01be5a42e4
Use more recent sentencepiece tokenizer API ( #696 )
2025-06-22 13:52:30 -05:00
Sebastian Raschka
bcfdbd7008
Fix some wording issues in the notes ( #695 )
2025-06-22 13:46:16 -05:00
Sajjad Baloch
661a6e84ee
Fix: Typo in appendix_d.py comments. ( #682 )
...
* Fix: pkg/llms_from_scratch/appendix_d.py
* minor language typo fix
* fix 691
---------
Co-authored-by: PrinceSajjadHussain <PrinceSajjadHussain@users.noreply.github.com>
Co-authored-by: rasbt <mail@sebastianraschka.com>
2025-06-22 12:15:12 -05:00
casinca
564e986496
fix issue #664 - inverted token and pos emb layers ( #665 )
...
* fix inverted token and pos layers
* remove redundant code
---------
Co-authored-by: rasbt <mail@sebastianraschka.com>
2025-06-22 12:15:01 -05:00
Sebastian Raschka
0a2e8c39c4
Qwen3 KV cache ( #688 )
2025-06-21 17:34:39 -05:00
Daniel Kleine
14c054d36c
added pkg fixes ( #676 )
...
Co-authored-by: Sebastian Raschka <mail@sebastianraschka.com>
2025-06-21 16:07:50 -05:00
Sebastian Raschka
fdc3e1b701
Add GPT-2 KV cache to pkg ( #687 )
2025-06-21 12:29:04 -05:00
Sebastian Raschka
3be0f3202a
Llama 3 KV Cache ( #685 )
...
* Llama 3 KV Cache
* skip expensive tests on Gh actions
* Update __init__.py
2025-06-21 10:55:20 -05:00
Sebastian Raschka
c008f95072
Fix formatting in Qwen3 nb ( #680 )
...
* Fix formatting in Qwen3 nb
* upd
2025-06-20 07:28:27 -05:00
casinca
00b8c0a107
minor readability improvements ( #668 )
2025-06-19 18:56:49 -05:00
Daniel Kleine
15fa6a84f6
fixed plot_losses ( #677 )
2025-06-19 18:55:43 -05:00
Sebastian Raschka
e719bd86ad
Qwen3 From Scratch ( #678 )
...
* Qwen3 From Scratch
* rev other file
* upd
* upd
* upd
* url fixes
2025-06-19 18:44:38 -05:00
casinca
58b8672452
removed old args in GQA class ( #674 )
2025-06-17 13:09:53 -05:00
Sebastian Raschka
ece59ba587
Optimize KV cache ( #673 )
...
* Optimize KV cache
* style
* interpretable generate
* interpretable generate
* update readme
2025-06-16 16:00:50 -05:00
Sebastian Raschka
ba0370abd1
Optimized KV cache ( #672 )
...
* Optimized KV cache
* typo fix
2025-06-15 14:26:16 -05:00
Sebastian Raschka
2af686d70b
Add KV cache ( #671 )
2025-06-15 09:58:08 -05:00
Sebastian Raschka
78bbcb3643
Remove redundant model = ( #663 )
2025-06-13 11:03:55 -05:00
Sebastian Raschka
3dfd7e5f06
Update pixi ( #661 )
...
* Llama3 from scratch improvements
* Update HF hub version in pixi.toml
* Update README.md
2025-06-13 10:50:17 -05:00
Daniel Kleine
c2cfb47b1a
fixed gqa qkv code comments ( #660 )
2025-06-13 08:21:28 -05:00
Greg Gandenberger
7632eb018b
Update ch07.ipynb ( #643 )
...
Correct function name
2025-06-13 08:17:10 -05:00
Shimpei Kojio
baaa6c9283
fixed video link ( #646 )
2025-06-13 08:16:18 -05:00
Pratyush Subhadarshi
d56417c34c
Correcting the wrong reference ( #649 )
...
Co-authored-by: Sebastian Raschka <mail@sebastianraschka.com>
2025-06-12 16:35:51 -05:00
Sebastian Raschka
c4cde1c21b
Reduce Llama 3 RoPE memory requirements ( #658 )
...
* Llama3 from scratch improvements
* Fix Llama 3 expensive RoPE memory issue
* updates
* update package
* benchmark
* remove unused rescale_theta
2025-06-12 11:08:02 -05:00
Sebastian Raschka
c278745aff
DeBERTa-v3 baseline ( #630 )
...
* Llama3 from scratch improvements
* deberta-baseline
* restore
2025-04-19 21:16:17 -05:00
Sebastian Raschka
4ff743051e
BPE cosmetics ( #629 )
...
* Llama3 from scratch improvements
* Cosmetic BPE improvements
* restore
* Update ch02/05_bpe-from-scratch/bpe-from-scratch.ipynb
* Update ch02/05_bpe-from-scratch/bpe-from-scratch.ipynb
* endoftext whitespace
2025-04-18 18:57:09 -05:00
Sebastian Raschka
adaf4faaae
Dpo vocab size clarification ( #628 )
...
* Llama3 from scratch improvements
* vocab size should be 50257 not 50256
* restore
2025-04-18 17:20:56 -05:00
Sebastian Raschka
47c036058d
Llama3 from scratch improvements ( #621 )
...
* Llama3 from scratch improvements
* restore
2025-04-16 18:08:26 -05:00
casinca
1b242d01a5
Minor DPO fixes ( #617 )
...
* minor dpo fixes
* Update dpo-from-scratch.ipynb
metadata diff
2025-04-16 12:56:49 -05:00
Daniel Kleine
f3d1566c2e
fixed <|endoftext|> token ( #620 )
2025-04-16 12:15:59 -05:00
Henry Shi
02779f5e35
updated exercise 5.3 ( #615 )
...
* updated exercise 5.3
temperature can be set to 0 to regardless of top_k setting to force deterministic behavior
* fix notebook json
---------
Co-authored-by: Sebastian Raschka <mail@sebastianraschka.com>
2025-04-13 13:06:57 -05:00
Sebastian Raschka
72efebd7f8
add special token handling to bpe from scratch code ( #616 )
2025-04-13 12:38:22 -05:00
Sebastian Raschka
92b308e512
Ch06 and Ch07 videos ( #613 )
...
* Ch06 and Ch07 videos
* exclude google scholar from link checking
2025-04-12 14:51:02 -05:00
PRASHANTH REDDY NIMMAKAYALA
74b9f1fcde
fix: typo in ch07.ipynb ( #612 )
2025-04-12 10:29:53 -05:00
Sebastian Raschka
9df572fdf4
Improve ModernBERT comments ( #606 )
...
* Improve modernbert comments
* bash code formatting
2025-04-06 18:29:22 -05:00
Sebastian Raschka
3654571184
align formulas in notes with code ( #605 )
2025-04-06 16:46:53 -05:00
Sebastian Raschka
67e0680210
Disable mask saving as weight in Llama 3 model ( #604 )
...
* Disable mask saving as weight
* update pixi
* update pixi
2025-04-06 09:33:36 -05:00
Sebastian Raschka
f1434652f2
reformat nbs ( #602 )
2025-04-05 16:18:27 -05:00
Sebastian Raschka
371ab9e8ff
Correct BERT experiments ( #600 )
2025-04-05 10:05:15 -05:00
Sebastian Raschka
4a9654173c
Add ModernBERT ( #598 )
2025-04-05 09:13:30 -05:00
Sebastian Raschka
d4c8d8f2c9
Fix Llama language typo in bonus materials ( #597 )
2025-04-02 21:41:36 -05:00
Sebastian Raschka
49330d0990
Fix link ( #596 )
2025-04-02 09:47:07 -05:00
Sebastian Raschka
43e25a5165
Llama3Fast ( #593 )
...
* Llama3Fast
* Update pkg/llms_from_scratch/tests/test_llama3.py
2025-04-01 12:56:11 -05:00
Sebastian Raschka
aedad7efc3
Add Llama 3.2 to pkg ( #591 )
...
* Add Llama 3.2 to pkg
* remove redundant attributes
* update tests
* updates
* updates
* updates
* fix link
* fix link
2025-03-31 18:59:47 -05:00
casinca
152a087a37
removing unused RoPE parameters ( #590 )
...
* removing unused RoPE parameters
* remove redundant context_length in GQA
---------
Co-authored-by: Sebastian Raschka <mail@sebastianraschka.com>
2025-03-31 17:10:39 -05:00
Sebastian Raschka
222803737d
Fix data download if UCI is temporarily down ( #592 )
2025-03-31 16:25:53 -05:00
Sebastian Raschka
6ea4dd3ae7
Clarify dataset length in chapter 2 ( #589 )
2025-03-30 16:01:37 -05:00
Sebastian Raschka
0f6894f41e
Memory optimized Llama ( #588 )
...
* Memory optimized Llama
* re-ad login
2025-03-30 15:18:12 -05:00
Sebastian Raschka
3f93d73d6d
Alt weight loading code via PyTorch ( #585 )
...
* Alt weight loading code via PyTorch
* commit additional files
2025-03-27 20:10:23 -05:00
