Commit graph

29 commits

Author SHA1 Message Date
Quinn Slack f4432e1dba
treat stop as stop sequences, not exact tokens (#442)
The `stop` option to the generate API is a list of sequences that should cause generation to stop. Although these are commonly called "stop tokens", they do not necessarily correspond to LLM tokens (per the LLM's tokenizer). For example, if the caller sends a generate request with `"stop":["\n"]`, then generation should stop on any token containing `\n` (and trim `\n` from the output), not just if the token exactly matches `\n`. If `stop` were interpreted strictly as LLM tokens, then it would require callers of the generate API to know the LLM's tokenizer and enumerate many tokens in the `stop` list.

Fixes https://github.com/jmorganca/ollama/issues/295.
2023-08-30 11:53:42 -04:00
Bruce MacDonald 519f4d98ef
add embed docs for modelfile 2023-08-17 13:37:42 -04:00
Bruce MacDonald 53bc36d207
Update modelfile.md 2023-08-15 09:23:36 -03:00
Bruce MacDonald 9ae9a89883 Update modelfile.md 2023-08-14 16:26:53 -03:00
Michael Yang 3a05d3def7
Merge pull request #326 from asarturas/document-num-gqa-parameter
Document num_gqa parameter
2023-08-10 18:18:38 -07:00
Arturas Smorgun d9c2687fd0
document default num_gqa to 1, as it's applicable to most models
Co-authored-by: Michael Yang <mxyng@pm.me>
2023-08-11 01:29:40 +01:00
Arturas Smorgun c0e7a3b90e
Document num_gqa parameter
It is required to be adjusted for some models, see https://github.com/jmorganca/ollama/issues/320 for more context
2023-08-11 00:58:09 +01:00
Michael Yang 37c9a8eea9 add lora docs 2023-08-10 09:23:40 -07:00
Bruce MacDonald 43c40c500e add embed docs for modelfile 2023-08-09 16:14:58 -04:00
Bruce MacDonald c4861360ec remove embed docs 2023-08-09 16:14:19 -04:00
Bruce MacDonald 1bee2347be pr feedback
- defer closing llm on embedding
- do not override licenses
- remove debugging print line
- reformat model file docs
2023-08-08 17:01:37 -04:00
Bruce MacDonald 3ceac05108 Add embedding docs 2023-08-08 14:04:11 -04:00
Michael Yang da2459d519
Update README.md (#265) 2023-08-02 22:38:32 -04:00
Bruce MacDonald 47bda0b860 add stop to docs 2023-07-28 12:30:27 -04:00
Bruce MacDonald fee7687cf3 Update modelfile.md 2023-07-27 15:15:10 -04:00
Bruce MacDonald 1ac38ec89c improve modelfile docs 2023-07-27 15:13:04 -04:00
Ikko Eltociear Ashimine 65d93a86b2
Update modelfile.md (#177)
fix markdown.
2023-07-22 08:19:30 -07:00
Jeffrey Morgan 31f0cb7742 new Modelfile syntax 2023-07-20 07:52:24 -07:00
Jeffrey Morgan 8945b25484 new modelfile syntax on branch 2023-07-20 02:24:21 -07:00
Michael Chiang 7c6ea2a966 fix dangling """ 2023-07-19 23:24:32 -07:00
Michael Chiang c161aef5f9 update example 2023-07-19 23:24:32 -07:00
Michael Chiang c47786c1b0 Update docs/modelfile.md
Co-authored-by: Michael Yang <mxyng@pm.me>
2023-07-19 23:24:32 -07:00
Michael Chiang df100ce540 Update docs/modelfile.md
Co-authored-by: Michael Yang <mxyng@pm.me>
2023-07-19 23:24:32 -07:00
Michael Chiang 5c5948b4e7 clean up my previous empty sentences 2023-07-19 23:24:32 -07:00
Michael Yang 1c72e46e09 update modelfile.md 2023-07-19 23:24:32 -07:00
Michael Chiang 25f874c030
Update modelfile.md 2023-07-19 12:48:57 -07:00
Michael Chiang 7fbdc8e2c1
Update modelfile.md 2023-07-19 11:38:06 -07:00
Matt Williams 67c8d49757 Updated modelfile doc to include license
and attributed midjourneyprompt

Signed-off-by: Matt Williams <m@technovangelist.com>
2023-07-19 07:16:38 -07:00
Matt Williams 9310ee3967 First stab at a modelfile doc
Signed-off-by: Matt Williams <m@technovangelist.com>
2023-07-18 08:22:17 -07:00