github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #6111] Request: add octopus-v4 #29578

New Issue
Open
opened 2026-04-22 08:34:24 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-22 08:34:24 -05:00
Owner
Copy Link

Originally created by @mak448a on GitHub (Aug 1, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/6111

Could you add octopus v4? (I haven't tested for malware, but it looks somewhat legit)

https://huggingface.co/NexaAIDev/Octopus-v4
https://huggingface.co/NexaAIDev/octopus-v4-gguf

Originally created by @mak448a on GitHub (Aug 1, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/6111 Could you add octopus v4? (I haven't tested for malware, but it looks somewhat legit) https://huggingface.co/NexaAIDev/Octopus-v4 https://huggingface.co/NexaAIDev/octopus-v4-gguf
GiteaMirror added the model label 2026-04-22 08:34:24 -05:00
GiteaMirror commented 2026-04-22 08:34:26 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Sep 11, 2024):

Ollama supports converting Phi-3-mini and Phi-3-medium out of the box. I believe Octopus v4 is based on Phi-3-mini, but uses a different template (it looks the same as Zephyr from their example card). I tried converting but ran into:

% ~/git/ollama/ollama create pdevine/octopus-v4
transferring model data 100%
converting model
creating new layer sha256:8fc27d21a13741f14883498037998588b1ab55ecb03bdb3ac0cbc47fc54c13f7
creating new layer sha256:2f877f969346ff11f1c4df30c717db8a797a6353a9a32e4e4b33f0a89ba81f5d
creating new layer sha256:b9fe8949313d978079167065437123c7603d980b0c3bba2214bf97296004f54c
creating new layer sha256:dc3f05170ba8dd9fba63b43a22dba06c3cb5099f8ff1d1a96cdfa6613c63b5a2
writing manifest
success

Here's the Modelfile I used:

FROM .
TEMPLATE """{{- range .Messages }}<|{{ .Role }}|>
{{ .Content }}</s>
{{ end }}<|assistant|>"""

PARAMETER stop <|system|>
PARAMETER stop </s>
PARAMETER stop <|user|>
PARAMETER stop <|assistant|>

Unfortunately here was the output I got:

>>> hi there
[UNK_BYTE_0xe29681▁It]It[UNK_BYTE_0xe29681▁appears]appears[UNK_BYTE_0xe29681▁there]there[UNK_BYTE_0xe29681▁was]was[UNK_BYTE_0xe29681▁an]an[UNK_BYTE_0xe29681▁error][UNK_BYTE_0xe29681▁It]It[UNK_BYTE_0xe29681▁appears]appears[UNK_BYTE_0xe29681▁there]there[UNK_BYTE_0xe29681▁was]was[UNK_BYTE_0xe29681▁an]an[UNK_BYTE_0xe29681▁error]error[UNK_BYTE_0xe29681▁in]in[UNK_BYTE_0xe29681▁the]the[UNK_BYTE_0xe29681▁input]input[UNK_BYTE_0xe29681▁prompt]prompt,[UNK_BYTE_0xe29681▁as]as[UNK_BYTE_0xe29681▁"]"rror[UNK_BYTE_0xe29681▁in]in[UNK_BYTE_0xe29681▁the]the[UNK_BYTE_0xe29681▁input]input[UNK_BYTE_0xe29681▁prompt]prompt,[UNK_BYTE_0xe29681▁as]as[UNK_BYTE_0xe29681▁"]"hi there"[UNK_BYTE_0xe29681▁does]does[UNK_BYTE_0xe29681▁not]not[UNK_BYTE_0xe29681▁form]form[UNK_BYTE_0xe29681▁a]a[UNK_BYTE_0xe29681▁coh]coherent[UNK_BYTE_0xe29681▁qthere"[UNK_BYTE_0xe29681▁does]does[UNK_BYTE_0xe29681▁not]not[UNK_BYTE_0xe29681▁form]form[UNK_BYTE_0xe29681▁a]a[UNK_BYTE_0xe29681▁coh]coherent[UNK_BYTE_0xe29681▁question]question[UNK_BYTE_0xe29681▁or]or[UNK_BYTE_0xe29681▁statement]statement.[UNK_BYTE_0xe29681▁Could]Could[UNK_BYTE_0xe29681▁you]you[UNK_BYTE_0xe29681▁please]pleestion]question[UNK_BYTE_0xe29681▁or]or[UNK_BYTE_0xe29681▁statement]statement.[UNK_BYTE_0xe29681▁Could]Could[UNK_BYTE_0xe29681▁you]you[UNK_BYTE_0xe29681▁please]please[UNK_BYTE_0xe29681▁provide]provide[UNK_BYTE_0xe29681▁additional]additional[UNK_BYTE_0xe29681▁context]context[UNK_BYTE_0xe29681▁or]or[UNK_BYTE_0xe29681▁re]rephrasse[UNK_BYTE_0xe29681▁provide]provide[UNK_BYTE_0xe29681▁additional]additional[UNK_BYTE_0xe29681▁context]context[UNK_BYTE_0xe29681▁or]or[UNK_BYTE_0xe29681▁re]rephrase[UNK_BYTE_0xe29681▁your]your[UNK_BYTE_0xe29681▁query]query?

I'm not sure what they're doing for a tokenizer, or if there's other parameters I need to tweak. The model card on HF wasn't super informative here.

<!-- gh-comment-id:2344433475 --> @pdevine commented on GitHub (Sep 11, 2024): Ollama supports converting Phi-3-mini and Phi-3-medium out of the box. I believe Octopus v4 is based on Phi-3-mini, but uses a different template (it looks the same as Zephyr from their example card). I tried converting but ran into: ``` % ~/git/ollama/ollama create pdevine/octopus-v4 transferring model data 100% converting model creating new layer sha256:8fc27d21a13741f14883498037998588b1ab55ecb03bdb3ac0cbc47fc54c13f7 creating new layer sha256:2f877f969346ff11f1c4df30c717db8a797a6353a9a32e4e4b33f0a89ba81f5d creating new layer sha256:b9fe8949313d978079167065437123c7603d980b0c3bba2214bf97296004f54c creating new layer sha256:dc3f05170ba8dd9fba63b43a22dba06c3cb5099f8ff1d1a96cdfa6613c63b5a2 writing manifest success ``` Here's the `Modelfile` I used: ``` FROM . TEMPLATE """{{- range .Messages }}<|{{ .Role }}|> {{ .Content }}</s> {{ end }}<|assistant|>""" PARAMETER stop <|system|> PARAMETER stop </s> PARAMETER stop <|user|> PARAMETER stop <|assistant|> ``` Unfortunately here was the output I got: ``` >>> hi there [UNK_BYTE_0xe29681▁It]It[UNK_BYTE_0xe29681▁appears]appears[UNK_BYTE_0xe29681▁there]there[UNK_BYTE_0xe29681▁was]was[UNK_BYTE_0xe29681▁an]an[UNK_BYTE_0xe29681▁error][UNK_BYTE_0xe29681▁It]It[UNK_BYTE_0xe29681▁appears]appears[UNK_BYTE_0xe29681▁there]there[UNK_BYTE_0xe29681▁was]was[UNK_BYTE_0xe29681▁an]an[UNK_BYTE_0xe29681▁error]error[UNK_BYTE_0xe29681▁in]in[UNK_BYTE_0xe29681▁the]the[UNK_BYTE_0xe29681▁input]input[UNK_BYTE_0xe29681▁prompt]prompt,[UNK_BYTE_0xe29681▁as]as[UNK_BYTE_0xe29681▁"]"rror[UNK_BYTE_0xe29681▁in]in[UNK_BYTE_0xe29681▁the]the[UNK_BYTE_0xe29681▁input]input[UNK_BYTE_0xe29681▁prompt]prompt,[UNK_BYTE_0xe29681▁as]as[UNK_BYTE_0xe29681▁"]"hi there"[UNK_BYTE_0xe29681▁does]does[UNK_BYTE_0xe29681▁not]not[UNK_BYTE_0xe29681▁form]form[UNK_BYTE_0xe29681▁a]a[UNK_BYTE_0xe29681▁coh]coherent[UNK_BYTE_0xe29681▁qthere"[UNK_BYTE_0xe29681▁does]does[UNK_BYTE_0xe29681▁not]not[UNK_BYTE_0xe29681▁form]form[UNK_BYTE_0xe29681▁a]a[UNK_BYTE_0xe29681▁coh]coherent[UNK_BYTE_0xe29681▁question]question[UNK_BYTE_0xe29681▁or]or[UNK_BYTE_0xe29681▁statement]statement.[UNK_BYTE_0xe29681▁Could]Could[UNK_BYTE_0xe29681▁you]you[UNK_BYTE_0xe29681▁please]pleestion]question[UNK_BYTE_0xe29681▁or]or[UNK_BYTE_0xe29681▁statement]statement.[UNK_BYTE_0xe29681▁Could]Could[UNK_BYTE_0xe29681▁you]you[UNK_BYTE_0xe29681▁please]please[UNK_BYTE_0xe29681▁provide]provide[UNK_BYTE_0xe29681▁additional]additional[UNK_BYTE_0xe29681▁context]context[UNK_BYTE_0xe29681▁or]or[UNK_BYTE_0xe29681▁re]rephrasse[UNK_BYTE_0xe29681▁provide]provide[UNK_BYTE_0xe29681▁additional]additional[UNK_BYTE_0xe29681▁context]context[UNK_BYTE_0xe29681▁or]or[UNK_BYTE_0xe29681▁re]rephrase[UNK_BYTE_0xe29681▁your]your[UNK_BYTE_0xe29681▁query]query? ``` I'm not sure what they're doing for a tokenizer, or if there's other parameters I need to tweak. The model card on HF wasn't super informative here.
GiteaMirror commented 2026-04-22 08:34:26 -05:00
Author
Owner
Copy Link

@CjhHa1 commented on GitHub (Sep 12, 2024):

Actually I have start an issue in their repo https://github.com/NexaAI/octopus-v4/issues/13
and @pdevine can you kindly explain why my error information is that Error: llama runner process has terminated: error loading modelvocabulary: _Map_base::at when I try to run

<!-- gh-comment-id:2345578209 --> @CjhHa1 commented on GitHub (Sep 12, 2024): Actually I have start an issue in their repo https://github.com/NexaAI/octopus-v4/issues/13 and @pdevine can you kindly explain why my error information is that `Error: llama runner process has terminated: error loading modelvocabulary: _Map_base::at` when I try to run
GiteaMirror commented 2026-04-22 08:34:27 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Sep 12, 2024):

@CjhHa1 I went ahead and downloaded one of their gguf files and tried it out. It looks like they created the GGUF files incorrectly.

<!-- gh-comment-id:2347248670 --> @pdevine commented on GitHub (Sep 12, 2024): @CjhHa1 I went ahead and downloaded one of their gguf files and tried it out. It looks like they created the GGUF files incorrectly.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label model
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#29578
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?