convert safetensor adapters into GGUF #6327
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
mxyng
requested changes
Aug 12, 2024
| // Convert writes an Ollama compatible model to the provided io.WriteSeeker based on configurations | ||
| // and files it finds in the input path. | ||
| // Supported input model formats include safetensors. | ||
| // Supported input tokenizers files include tokenizer.json (preferred) and tokenizer.model. | ||
| func Convert(fsys fs.FS, ws io.WriteSeeker) error { | ||
| func ConvertModel(fsys fs.FS, ws io.WriteSeeker) error { |
There was a problem hiding this comment.
Convert should remain the same but call convertAdapter() internally
08dd531 to
e7c08fe
Compare
mxyng
reviewed
Aug 13, 2024
convert/convert.go
Outdated
| KV(llm.KV) llm.KV | ||
| Tensors([]Tensor) []llm.Tensor | ||
|
|
||
| tensorName(string) string |
There was a problem hiding this comment.
#5365 changes how tensor names are mapped so be aware
943f3f3 to
da3d6a2
Compare
jmorganca
reviewed
Aug 19, 2024
|
|
||
| return conv.writeFile(ws, conv.KV(baseKV), conv.Tensors(ts)) | ||
| } | ||
|
|
||
| // Convert writes an Ollama compatible model to the provided io.WriteSeeker based on configurations |
There was a problem hiding this comment.
Comment needs an update if we're going to change this
jmorganca
reviewed
Aug 19, 2024
| @@ -11,12 +11,21 @@ import ( | |||
| "github.com/ollama/ollama/llm" | |||
| ) | |||
|
|
|||
| type Parameters struct { | |||
| type ModelParameters struct { | |||
There was a problem hiding this comment.
it's implicit (as it is elsewhere in the product) that we're dealing with models already
Suggested change
| type ModelParameters struct { | |
| type Parameters struct { |
jmorganca
approved these changes
Aug 19, 2024
da3d6a2 to
d722cd5
Compare
d722cd5 to
8360dc0
Compare
mxyng
reviewed
Aug 22, 2024
mxyng
approved these changes
Aug 22, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This change converts a Safetensors based LoRA into GGUF and ties it w/ a base model. Only llama2/llama3/mistral/gemma2 will work initially. You can create the Modelfile to look like:
I'll add in some tests, but wanted to get this out so people could try it out.
Replaces #5524