Difference between revisions of "LLM: tips untuk CPU"
Jump to navigation
Jump to search
Onnowpurbo (talk | contribs) (Created page with "Kata CGPT: saat pake CPU, coba: 1. Batch Processing u. kurangi overhead & speedup embedding. 2. Kurangi presisi model; float32->float16/int8; speedup tanpa korbankan akuras...") |
Onnowpurbo (talk | contribs) |
||
| (2 intermediate revisions by the same user not shown) | |||
| Line 5: | Line 5: | ||
4. Multi-threading. | 4. Multi-threading. | ||
5. Gunakan Intel MKL / OpenBLAS. | 5. Gunakan Intel MKL / OpenBLAS. | ||
| + | |||
| + | saya pakai model intfloat pak, lumayan cepet di CPU https://huggingface.co/intfloat/multilingual-e5-large | ||
| + | |||
| + | |||
| + | |||
| + | kalo pdf bisanya saya parse dulu textnya atau pakai ocr, terus embeddingnya disimpan di postgre pakai pgvector (https://github.com/pgvector/pgvector) | ||
| + | agak effort sih | ||
Latest revision as of 04:26, 17 July 2024
Kata CGPT: saat pake CPU, coba:
1. Batch Processing u. kurangi overhead & speedup embedding. 2. Kurangi presisi model; float32->float16/int8; speedup tanpa korbankan akurasi. 3. Buat versi kecil dari model yg sama. 4. Multi-threading. 5. Gunakan Intel MKL / OpenBLAS.
saya pakai model intfloat pak, lumayan cepet di CPU https://huggingface.co/intfloat/multilingual-e5-large
kalo pdf bisanya saya parse dulu textnya atau pakai ocr, terus embeddingnya disimpan di postgre pakai pgvector (https://github.com/pgvector/pgvector) agak effort sih