Ollama （0.4.7） 加载失败

[zhcharles]

System Info / 系統信息

Rockylinux 9

Who can help? / 谁可以帮助到您？

No response

Information / 问题信息

• The official example scripts / 官方的示例脚本
• My own modified scripts / 我自己修改的脚本和任务

Reproduction / 复现过程

ollama run hf.co/THUDM/glm-edge-1.5b-chat-gguf:Q4_0

pulling manifest
pulling manifest
pulling manifest
pulling db55499d77d3... 100% ▕███████████████████████████████▏ 931 MB
pulling 710a4abbfbd3... 100% ▕████████████████████████████████▏ 6.6 KB
pulling af0ddbdaaa26... 100% ▕████████████████████████████████▏ 70 B
pulling 66dfcc32a723... 100% ▕████████████████████████████████▏ 57 B
pulling 16e6f92d5b71... 100% ▕████████████████████████████████
verifying sha256 digest
writing manifest
success
Error: llama runner process has terminated: error loading model: check_tensor_dims: tensor 'blk.0.attn_qkv.weight' not found
llama_load_model_from_file: failed to load model

在 HF 上右上角，可以获得 ollama 的下载命令，下载成功。但是 ollama run 时，出现以上问题。

Expected behavior / 期待表现

正确加载并使用

[wukonggeo]

@zRzRzRzRzRzRzR 不支持ollama部署么？

[minghao51]

it somehow works for me on later ollama version,
perhaps update your ollama?

[piDack]

ggml-org/llama.cpp#10573 llama.cpp merge in Feb. I think it work great in ollama. this issue may be closed @zRzRzRzRzRzRzR