github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-08 08:59:13 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #13888] Using previously downloaded models in .15 no longer work. #71148

New Issue
Closed
opened 2026-05-05 00:33:02 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-05-05 00:33:02 -05:00
Owner
Copy Link

Originally created by @sammyvoncheese on GitHub (Jan 24, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/13888

What is the issue?

Error loading any previously downloaded models in .15. Getting this error calling the python API against the .15 server. Does not occur in previous version.

Example Model: glm-4.7-flash:Q4_K_M

Redownloading the model allows it to work. I have 228 models downloaded already; redownloading terabytes of models is a non-starter.

Windows 11, binaries install only.

I tried using the newly re-downloaded model glm-4.7-flash:Q4_K_M against .14.3 and got this

time=2026-01-24T15:48:17.478-05:00 level=INFO source=server.go:3634 msg="http: panic serving 127.0.0.1:60938: runtime error: invalid memory address or nil pointer dereference\ngoroutine 52 [running]:\nnet/http.(*conn).serve.func1()\n\tnet/http/server.go:1947 +0xbe\npanic({0x7ff79e4e1f20?, 0x7ff79f207790?})\n\truntime/panic.go:792 +0x132\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).allocModel.func1()\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1187 +0x11a\npanic({0x7ff79e4e1f20?, 0x7ff79f207790?})\n\truntime/panic.go:792 +0x132\ngithub.com/ollama/ollama/ml/nn.(*Linear).Forward(0x0, {0x7ff79e859340, 0xc00148ac00}, {0x7ff79e867680?, 0xc00c8801e0?})\n\tgithub.com/ollama/ollama/ml/nn/linear.go:11 +0x27\ngithub.com/ollama/ollama/model/models/glm4moelite.(*Attention).Forward(0xc001076400, {0x7ff79e859340, 0xc00148ac00}, {0x7ff79e867680, 0xc00c880090}, {0x7ff79e867680, 0xc00c880060}, {0x7ff79e855658, 0xc001354100}, 0xc000525500)\n\tgithub.com/ollama/ollama/model/models/glm4moelite/model.go:81 +0x634\ngithub.com/ollama/ollama/model/models/glm4moelite.(*Layer).Forward(0xc0000472d8, {0x7ff79e859340, 0xc00148ac00}, {0x7ff79e867680, 0xc00c880078}, {0x7ff79e867680, 0xc00c880060}, {0x0, 0x0}, {0x7ff79e855658, ...}, ...)\n\tgithub.com/ollama/ollama/model/models/glm4moelite/model.go:177 +0xd9\ngithub.com/ollama/ollama/model/models/glm4moelite.(*Model).Forward(0xc0000d8b80, {0x7ff79e859340, 0xc00148ac00}, {{0x7ff79e867680, 0xc001494798}, {0x7ff79e867680, 0xc0014947b0}, {0xc0014a4800, 0x200, 0x200}, ...})\n\tgithub.com/ollama/ollama/model/models/glm4moelite/model.go:295 +0x16a\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).reserveWorstCaseGraph(0xc00022b0e0, 0x1)\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1157 +0x9ad\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).allocModel(0xc00022b0e0, {0xc00002c060?, 0x7ff79d54af9a?}, {0x0, 0x10, {0xc0001441c0, 0x1, 0x1}, 0x1}, {0x0, ...}, ...)\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1226 +0x391\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).load(0xc00022b0e0, {0x7ff79e84b060, 0xc0001640e0}, 0xc00014e280)\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1305 +0x54b\nnet/http.HandlerFunc.ServeHTTP(0xc0006a8780?, {0x7ff79e84b060?, 0xc0001640e0?}, 0xc000047b60?)\n\tnet/http/server.go:2294 +0x29\nnet/http.(*ServeMux).ServeHTTP(0x7ff79d1eb785?, {0x7ff79e84b060, 0xc0001640e0}, 0xc00014e280)\n\tnet/http/server.go:2822 +0x1c4\nnet/http.serverHandler.ServeHTTP({0x7ff79e847570?}, {0x7ff79e84b060?, 0xc0001640e0?}, 0x1?)\n\tnet/http/server.go:3301 +0x8e\nnet/http.(*conn).serve(0xc0006e2480, {0x7ff79e84d4c8, 0xc0006def60})\n\tnet/http/server.go:2102 +0x625\ncreated by net/http.(*Server).Serve in goroutine 1\n\tnet/http/server.go:3454 +0x485"

Relevant log output

Error: failed to initialize model: this model uses a weight format that is no longer supported; please re-download it
(status code: 500)

OS

Windows 11

GPU

5090

CPU

AMD Ryzen 9, 995x03D 16 Core

Ollama version

.15

Originally created by @sammyvoncheese on GitHub (Jan 24, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/13888 ### What is the issue? Error loading any previously downloaded models in .15. Getting this error calling the python API against the .15 server. Does not occur in previous version. Example Model: glm-4.7-flash:Q4_K_M Redownloading the model allows it to work. I have 228 models downloaded already; redownloading terabytes of models is a non-starter. Windows 11, binaries install only. I tried using the newly re-downloaded model glm-4.7-flash:Q4_K_M against .14.3 and got this time=2026-01-24T15:48:17.478-05:00 level=INFO source=server.go:3634 msg="http: panic serving 127.0.0.1:60938: runtime error: invalid memory address or nil pointer dereference\ngoroutine 52 [running]:\nnet/http.(*conn).serve.func1()\n\tnet/http/server.go:1947 +0xbe\npanic({0x7ff79e4e1f20?, 0x7ff79f207790?})\n\truntime/panic.go:792 +0x132\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).allocModel.func1()\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1187 +0x11a\npanic({0x7ff79e4e1f20?, 0x7ff79f207790?})\n\truntime/panic.go:792 +0x132\ngithub.com/ollama/ollama/ml/nn.(*Linear).Forward(0x0, {0x7ff79e859340, 0xc00148ac00}, {0x7ff79e867680?, 0xc00c8801e0?})\n\tgithub.com/ollama/ollama/ml/nn/linear.go:11 +0x27\ngithub.com/ollama/ollama/model/models/glm4moelite.(*Attention).Forward(0xc001076400, {0x7ff79e859340, 0xc00148ac00}, {0x7ff79e867680, 0xc00c880090}, {0x7ff79e867680, 0xc00c880060}, {0x7ff79e855658, 0xc001354100}, 0xc000525500)\n\tgithub.com/ollama/ollama/model/models/glm4moelite/model.go:81 +0x634\ngithub.com/ollama/ollama/model/models/glm4moelite.(*Layer).Forward(0xc0000472d8, {0x7ff79e859340, 0xc00148ac00}, {0x7ff79e867680, 0xc00c880078}, {0x7ff79e867680, 0xc00c880060}, {0x0, 0x0}, {0x7ff79e855658, ...}, ...)\n\tgithub.com/ollama/ollama/model/models/glm4moelite/model.go:177 +0xd9\ngithub.com/ollama/ollama/model/models/glm4moelite.(*Model).Forward(0xc0000d8b80, {0x7ff79e859340, 0xc00148ac00}, {{0x7ff79e867680, 0xc001494798}, {0x7ff79e867680, 0xc0014947b0}, {0xc0014a4800, 0x200, 0x200}, ...})\n\tgithub.com/ollama/ollama/model/models/glm4moelite/model.go:295 +0x16a\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).reserveWorstCaseGraph(0xc00022b0e0, 0x1)\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1157 +0x9ad\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).allocModel(0xc00022b0e0, {0xc00002c060?, 0x7ff79d54af9a?}, {0x0, 0x10, {0xc0001441c0, 0x1, 0x1}, 0x1}, {0x0, ...}, ...)\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1226 +0x391\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).load(0xc00022b0e0, {0x7ff79e84b060, 0xc0001640e0}, 0xc00014e280)\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1305 +0x54b\nnet/http.HandlerFunc.ServeHTTP(0xc0006a8780?, {0x7ff79e84b060?, 0xc0001640e0?}, 0xc000047b60?)\n\tnet/http/server.go:2294 +0x29\nnet/http.(*ServeMux).ServeHTTP(0x7ff79d1eb785?, {0x7ff79e84b060, 0xc0001640e0}, 0xc00014e280)\n\tnet/http/server.go:2822 +0x1c4\nnet/http.serverHandler.ServeHTTP({0x7ff79e847570?}, {0x7ff79e84b060?, 0xc0001640e0?}, 0x1?)\n\tnet/http/server.go:3301 +0x8e\nnet/http.(*conn).serve(0xc0006e2480, {0x7ff79e84d4c8, 0xc0006def60})\n\tnet/http/server.go:2102 +0x625\ncreated by net/http.(*Server).Serve in goroutine 1\n\tnet/http/server.go:3454 +0x485" ### Relevant log output Error: failed to initialize model: this model uses a weight format that is no longer supported; please re-download it (status code: 500) ### OS Windows 11 ### GPU 5090 ### CPU AMD Ryzen 9, 995x03D 16 Core ### Ollama version .15
GiteaMirror added the bug label 2026-05-05 00:33:02 -05:00
GiteaMirror commented 2026-05-05 00:33:15 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 24, 2026):

Only glm-4.7-flash needs to be refreshed for the weight format change, 0.15.0 implements a different weight to reduce the memory footprint. The crash (invalid memory address or nil pointer dereference) should be unrelated to that. A full log after refreshing the model would be helpful.

<!-- gh-comment-id:3795562188 --> @rick-github commented on GitHub (Jan 24, 2026): Only glm-4.7-flash needs to be refreshed for the weight format change, 0.15.0 implements a different weight to reduce the memory footprint. The crash (`invalid memory address or nil pointer dereference`) should be unrelated to that. A full log after refreshing the model would be helpful.
GiteaMirror commented 2026-05-05 00:33:17 -05:00
Author
Owner
Copy Link

@UltraRabbit commented on GitHub (Jan 25, 2026):

I tried to use 'ollam pull glm-4.7-flash:Q4_K_M' again when the old model still in my local. It does seem to download the new data but failed when launching 'ollama run glm-4.7-flash:Q4_K_M' and giving the same error message. After I remove the local model and repull it from ollama repo, it could be load and run without memory consumption issue, but still having thinking loop. However, after download the unsloth's gguf model from huggingface and imported into ollama, the thinking loop was fixed but memory consumption issue came back again.

<!-- gh-comment-id:3795858527 --> @UltraRabbit commented on GitHub (Jan 25, 2026): I tried to use 'ollam pull glm-4.7-flash:Q4_K_M' again when the old model still in my local. It does seem to download the new data but failed when launching 'ollama run glm-4.7-flash:Q4_K_M' and giving the same error message. After I remove the local model and repull it from ollama repo, it could be load and run without memory consumption issue, but still having thinking loop. However, after download the unsloth's gguf model from huggingface and imported into ollama, the thinking loop was fixed but memory consumption issue came back again.
GiteaMirror commented 2026-05-05 00:33:18 -05:00
Author
Owner
Copy Link

@sammyvoncheese commented on GitHub (Jan 25, 2026):

@rick-github the error was using the new model downloaded with .15 with the 14.3 server. I had deleted the old one (pre .15 not sure when I pulled it now), then downloaded the new one with .15.

Very cool that the new version runs 28gb on my 5090 with 90000 context, I think the last one was like 100 gigs on my machine.

<!-- gh-comment-id:3795868905 --> @sammyvoncheese commented on GitHub (Jan 25, 2026): @rick-github the error was using the new model downloaded with .15 with the 14.3 server. I had deleted the old one (pre .15 not sure when I pulled it now), then downloaded the new one with .15. Very cool that the new version runs 28gb on my 5090 with 90000 context, I think the last one was like 100 gigs on my machine.
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 bug
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#71148
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?