ggerganov/llama.cpp

Pull Requests Commits
(Windows) Set console to UTF-8 on init (#420)
anzz1 committed 2 years ago
Verified 34ab5268
Fix colors enabling on WIN32
ggerganov committed 2 years ago
Verified c2b25b69
If n_predict == -1, generate forever
ggerganov committed 2 years ago
Verified 79b2b266
Inifinite generation via context swapping (#71)
ggerganov committed 2 years ago
Verified e2d490da
Cleanup STL headers + fix embedding examples + minor stuff
ggerganov committed 2 years ago
Verified 03f7e335
Move chat scripts into "./examples"
ggerganov committed 2 years ago
Verified 55ad42af
Add AVX2 implementation of dequantize_row_q4_1 (#505)
slaren committed 2 years ago
Verified 459e93cc
Overhaul the examples structure
ggerganov committed 2 years ago
Verified a316a425
Retire the ggml_mul_mat() branch for transposed src0 (#500)
ggerganov committed 2 years ago
Verified ecbe466a
Disable prompt verbosity by default and add option to enable (#480)
ggerganov committed 2 years ago
Verified 502a4001
Add AVX2 implementation of dequantize_row_q4_0 (#467)
slaren committed 2 years ago
Verified 09aecbf6
Don't interefe with BLAS for large prompts by running only 1 thread
ggerganov committed 2 years ago
Verified 4640eff2
Add longer DAN prompt for testing big batch numbers
ggerganov committed 2 years ago
Verified ab77d763
Add timings for the prompt evaluation (#478)
slaren committed 2 years ago
Verified 29b7baab
Remove obsolete information from README
ggerganov committed 2 years ago
Verified 4a7129ac
Remove obsolete assert and fix compiler warning
ggerganov committed 2 years ago
Verified 6b6dbc89
Fix nasty bug in ggml_compute_forward_mul_mat_f32() and reenable BLAS
ggerganov committed 2 years ago
Verified 2a2e63ce
bounds checking for input prefix (#492)
anzz1 committed 2 years ago
Verified e899bf54
feat: '--in-prefix STRING' option (#426)
anzz1 committed 2 years ago
Verified fbd4d38c
Add support for file load progress reporting callbacks (#434)
j-f1 committed 2 years ago
Verified 58e6c9f3
Add missing struct annotation (#483)
Doomsdayrs committed 2 years ago
Verified 36d07532
Fix crash for 65B model with pre-allocated memory (#485)
chriskuehl committed 2 years ago
Verified 6f1ee4b6
Disable BLAS altogether - the bug is not just for qunatized mat mul
ggerganov committed 2 years ago
Verified 8520fc31
Disable BLAS branch in mul_mat - seems there is a bug
ggerganov committed 2 years ago
Verified b3f460e9
Immediately start processing the prompt before user input has been provided (#476)
ggerganov committed 2 years ago
Verified 04c6f5ed
Reduce memory usage and allocate enough memory for largest context (#473)
ggerganov committed 2 years ago
Verified 7a9b6c3a
Temporary bump the memory buffer size - hopefully fix issues from 483bab2e
ggerganov committed 2 years ago
Verified 31572d96
Update README.md (#444)
gjmulder committed 2 years ago
Verified f4f5362e
fix instruct mode (#445)
rabidcopy committed 2 years ago
Verified 863f65e2
Properly free llama_context on failure
ggerganov committed 2 years ago
Verified afd220d9
Terms Of Use Privacy Policy Refund Policy Impressum