tjohnman
|
ad5fd5b60c
Make prompt randomization optional. (#300)
|
2 лет назад |
tjohnman
|
368d0c8a9e
Respect the maximum number of tokens in interactive. (#298)
|
2 лет назад |
slaren
|
50fae10d03
Add --ignore-eos parameter (#181)
|
2 лет назад |
Qingyou Meng
|
084e2f0ec0
interactive mode: print '\n' in sigint_handler, this flush stdout thus ensure color reset. (#283)
|
2 лет назад |
Erik Scholz
|
0b366e7357
Command line switch to use F16 for memory_k and memory_v (refactor of #154) (#294)
|
2 лет назад |
Georgi Gerganov
|
c494ed5b94
Fix off-by-one bug (#115)
|
2 лет назад |
Georgi Gerganov
|
70f01cb863
Drop trailing new line from file prompts (#80)
|
2 лет назад |
Georgi Gerganov
|
9e1707218a
Add "--instruct" argument for usage with Alpaca (#240)
|
2 лет назад |
Ronsor
|
d7def1a752
Warn user if a context size greater than 2048 tokens is specified (#274)
|
2 лет назад |
Alex Nguyen
|
d3f202d57b
Remove unused code since n_vocab is model.hparams.n_vocab (#262)
|
2 лет назад |
Justin Suess
|
e03e359730
fixed warning with std::ignore about unused function result (#151)
|
2 лет назад |
thement
|
c9f670a177
Implement non-greedy tokenizer that tries to maximize token lengths (#242)
|
2 лет назад |
hoangmit
|
6eac39ba95
Add RMS norm and use it (#187)
|
2 лет назад |
Rickey Bowers Jr
|
2d15d6c9a9
add SIGINT support for _WIN32 environments (#120)
|
2 лет назад |
Justin Suess
|
2d64715ad4
added ctx_size parameter (#148)
|
2 лет назад |
Justin Suess
|
16b2c61a22
fixed color reset on exit (#149)
|
2 лет назад |
Georgi Gerganov
|
4497ad819c
Print system information
|
2 лет назад |
Pavol Rusnak
|
671d5cac15
Use fprintf for diagnostic output (#48)
|
2 лет назад |
uint256_t
|
63fd76fbb0
Reduce model loading time (#43)
|
2 лет назад |
Val Kharitonov
|
2a20f48efa
Fix UTF-8 handling (including colors) (#79)
|
2 лет назад |
Matvey Soloviev
|
a169bb889c
Gate signal support on being on a unixoid system. (#74)
|
2 лет назад |
Matvey Soloviev
|
460c482540
Fix token count accounting
|
2 лет назад |
Matvey Soloviev
|
404fac0d62
Fix color getting reset before prompt output done (#65)
|
2 лет назад |
Matvey Soloviev
|
96ea727f47
Add interactive mode (#61)
|
2 лет назад |
beiller
|
02f0c6fe7f
Add back top_k (#56)
|
2 лет назад |
Sebastián A
|
eb062bb012
Windows fixes (#31)
|
2 лет назад |
beiller
|
129c7d1ea8
Add repetition penalty (#20)
|
2 лет назад |
Georgi Gerganov
|
7d9ed7b25f
Bump memory buffer
|
2 лет назад |
Georgi Gerganov
|
007a8f6f45
Support all LLaMA models + change Q4_0 quantization storage
|
2 лет назад |
Georgi Gerganov
|
70bc0b8b15
Fix a bug in the rope calculation
|
2 лет назад |