github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #13529] loading models in old gpu with the new update is not working at all ( it was working well until i update it ) #55424

New Issue
Open
opened 2026-04-29 09:09:38 -05:00 by GiteaMirror · 12 comments
GiteaMirror commented 2026-04-29 09:09:38 -05:00
Owner
Copy Link

Originally created by @elsare7now on GitHub (Dec 19, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/13529

What is the issue?

500 Internal Server Error: llama runner process has terminated: error loading model: missing tensor 'output_norm'

Relevant log output

OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

No response

Originally created by @elsare7now on GitHub (Dec 19, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/13529 ### What is the issue? 500 Internal Server Error: llama runner process has terminated: error loading model: missing tensor 'output_norm' ### Relevant log output ```shell ``` ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version _No response_
GiteaMirror added the bug label 2026-04-29 09:09:38 -05:00
GiteaMirror commented 2026-04-29 09:09:41 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Dec 19, 2025):

Server log may help in debugging.

<!-- gh-comment-id:3675061792 --> @rick-github commented on GitHub (Dec 19, 2025): [Server log](https://docs.ollama.com/troubleshooting) may help in debugging.
GiteaMirror commented 2026-04-29 09:09:42 -05:00
Author
Owner
Copy Link

@qustrolabe commented on GitHub (Dec 20, 2025):

I have the same issue with sam860/LFM2:350m and sam860/LFM2:1.2b models, trying some other models I have and they work flawlessly. I remember running those 2 broken ones around month ago without any problem before recent update.

Win 11 ollama 0.13.5

Logs don't say much besides error loading model: missing tensor 'output_norm'

Logs 
time=2025-12-20T05:21:40.813+02:00 level=INFO source=routes.go:1554 msg="server config" env="map[CUDA_VISIBLE_DEVICES: GGML_VK_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:4096 OLLAMA_DEBUG:INFO OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:C:\\Users\\User\\.ollama\\models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_REMOTES:[ollama.com] OLLAMA_SCHED_SPREAD:false OLLAMA_VULKAN:false ROCR_VISIBLE_DEVICES:]"
time=2025-12-20T05:21:40.819+02:00 level=INFO source=images.go:493 msg="total blobs: 28"
time=2025-12-20T05:21:40.821+02:00 level=INFO source=images.go:500 msg="total unused blobs removed: 0"
time=2025-12-20T05:21:40.822+02:00 level=INFO source=routes.go:1607 msg="Listening on 127.0.0.1:11434 (version 0.13.5)"
time=2025-12-20T05:21:40.823+02:00 level=INFO source=runner.go:67 msg="discovering available GPUs..."
time=2025-12-20T05:21:40.840+02:00 level=INFO source=runner.go:106 msg="experimental Vulkan support disabled.  To enable, set OLLAMA_VULKAN=1"
time=2025-12-20T05:21:40.849+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 53677"
time=2025-12-20T05:21:41.040+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 53682"
time=2025-12-20T05:21:41.228+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 53689"
time=2025-12-20T05:21:41.327+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 53695"
time=2025-12-20T05:21:41.327+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 53696"
time=2025-12-20T05:21:41.535+02:00 level=INFO source=types.go:42 msg="inference compute" id=GPU-fb36d98c-caf3-51fa-1c88-c1a9ec3cf2e7 filter_id="" library=CUDA compute=7.5 name=CUDA0 description="NVIDIA GeForce GTX 1650" libdirs=ollama,cuda_v13 driver=13.0 pci_id=0000:01:00.0 type=discrete total="4.0 GiB" available="3.6 GiB"
time=2025-12-20T05:21:41.535+02:00 level=INFO source=routes.go:1648 msg="entering low vram mode" "total vram"="4.0 GiB" threshold="20.0 GiB"
[GIN] 2025/12/20 - 05:21:41 | 200 |            0s |       127.0.0.1 | GET      "/api/version"
[GIN] 2025/12/20 - 05:21:41 | 200 |            0s |       127.0.0.1 | HEAD     "/"
[GIN] 2025/12/20 - 05:21:41 | 200 |     24.5469ms |       127.0.0.1 | POST     "/api/show"
[GIN] 2025/12/20 - 05:21:41 | 200 |     24.4573ms |       127.0.0.1 | POST     "/api/show"
time=2025-12-20T05:21:41.623+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 53707"
time=2025-12-20T05:21:41.814+02:00 level=INFO source=cpu_windows.go:148 msg=packages count=1
time=2025-12-20T05:21:41.814+02:00 level=INFO source=cpu_windows.go:195 msg="" package=0 cores=4 efficiency=0 threads=8
llama_model_loader: loaded meta data with 34 key-value pairs and 148 tensors from C:\Users\User\.ollama\models\blobs\sha256-b7bfeab6495a1ae3ae78811c1297df9f301b35261ff9580d42fb30dc4dc9034b (version GGUF V3 (latest))
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv   0:                       general.architecture str              = lfm2
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = LFM2 350M
llama_model_loader: - kv   3:                           general.basename str              = LFM2
llama_model_loader: - kv   4:                         general.size_label str              = 350M
llama_model_loader: - kv   5:                            general.license str              = other
llama_model_loader: - kv   6:                       general.license.name str              = lfm1.0
llama_model_loader: - kv   7:                       general.license.link str              = LICENSE
llama_model_loader: - kv   8:                               general.tags arr[str,4]       = ["liquid", "lfm2", "edge", "text-gene...
llama_model_loader: - kv   9:                          general.languages arr[str,8]       = ["en", "ar", "zh", "fr", "de", "ja", ...
llama_model_loader: - kv  10:                           lfm2.block_count u32              = 16
llama_model_loader: - kv  11:                        lfm2.context_length u32              = 128000
llama_model_loader: - kv  12:                      lfm2.embedding_length u32              = 1024
llama_model_loader: - kv  13:                   lfm2.feed_forward_length u32              = 4608
llama_model_loader: - kv  14:                  lfm2.attention.head_count u32              = 16
llama_model_loader: - kv  15:               lfm2.attention.head_count_kv arr[i32,16]      = [0, 0, 8, 0, 0, 8, 0, 0, 8, 0, 8, 0, ...
llama_model_loader: - kv  16:                        lfm2.rope.freq_base f32              = 1000000.000000
llama_model_loader: - kv  17:                            lfm2.vocab_size u32              = 65536
llama_model_loader: - kv  18:                     lfm2.shortconv.l_cache u32              = 3
llama_model_loader: - kv  19:      lfm2.attention.layer_norm_rms_epsilon f32              = 0.000010
llama_model_loader: - kv  20:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  21:                         tokenizer.ggml.pre str              = lfm2
llama_model_loader: - kv  22:                      tokenizer.ggml.tokens arr[str,65536]   = ["<|pad|>", "<|startoftext|>", "<|end...
llama_model_loader: - kv  23:                  tokenizer.ggml.token_type arr[i32,65536]   = [3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, ...
llama_model_loader: - kv  24:                      tokenizer.ggml.merges arr[str,63683]   = ["Ċ Ċ", "Ċ ĊĊ", "ĊĊ Ċ", "Ċ Į..
llama_model_loader: - kv  25:                tokenizer.ggml.bos_token_id u32              = 1
llama_model_loader: - kv  26:                tokenizer.ggml.eos_token_id u32              = 7
llama_model_loader: - kv  27:            tokenizer.ggml.padding_token_id u32              = 0
llama_model_loader: - kv  28:               tokenizer.ggml.add_bos_token bool             = true
llama_model_loader: - kv  29:               tokenizer.ggml.add_sep_token bool             = false
llama_model_loader: - kv  30:               tokenizer.ggml.add_eos_token bool             = false
llama_model_loader: - kv  31:                    tokenizer.chat_template str              = {{bos_token}}{% for message in messag...
llama_model_loader: - kv  32:               general.quantization_version u32              = 2
llama_model_loader: - kv  33:                          general.file_type u32              = 7
llama_model_loader: - type  f32:   55 tensors
llama_model_loader: - type q8_0:   93 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type   = Q8_0
print_info: file size   = 359.37 MiB (8.50 BPW) 
load: printing all EOG tokens:
load:   - 2 ('<|endoftext|>')
load:   - 7 ('<|im_end|>')
load: special tokens cache size = 507
load: token to piece cache size = 0.3756 MB
print_info: arch             = lfm2
print_info: vocab_only       = 1
print_info: no_alloc         = 0
print_info: model type       = ?B
print_info: model params     = 354.48 M
print_info: general.name     = LFM2 350M
print_info: vocab type       = BPE
print_info: n_vocab          = 65536
print_info: n_merges         = 63683
print_info: BOS token        = 1 '<|startoftext|>'
print_info: EOS token        = 7 '<|im_end|>'
print_info: EOT token        = 2 '<|endoftext|>'
print_info: PAD token        = 0 '<|pad|>'
print_info: LF token         = 708 'Ċ'
print_info: EOG token        = 2 '<|endoftext|>'
print_info: EOG token        = 7 '<|im_end|>'
print_info: max token length = 30
llama_model_load: vocab only - skipping tensors
time=2025-12-20T05:21:41.912+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --model C:\\Users\\User\\.ollama\\models\\blobs\\sha256-b7bfeab6495a1ae3ae78811c1297df9f301b35261ff9580d42fb30dc4dc9034b --port 53712"
time=2025-12-20T05:21:41.916+02:00 level=INFO source=sched.go:443 msg="system memory" total="15.8 GiB" free="4.2 GiB" free_swap="6.8 GiB"
time=2025-12-20T05:21:41.916+02:00 level=INFO source=sched.go:450 msg="gpu memory" id=GPU-fb36d98c-caf3-51fa-1c88-c1a9ec3cf2e7 library=CUDA available="3.2 GiB" free="3.6 GiB" minimum="457.0 MiB" overhead="0 B"
time=2025-12-20T05:21:41.916+02:00 level=INFO source=server.go:496 msg="loading model" "model layers"=17 requested=-1
time=2025-12-20T05:21:41.916+02:00 level=INFO source=device.go:240 msg="model weights" device=CUDA0 size="359.4 MiB"
time=2025-12-20T05:21:41.916+02:00 level=INFO source=device.go:251 msg="kv cache" device=CUDA0 size="48.0 MiB"
time=2025-12-20T05:21:41.916+02:00 level=INFO source=device.go:262 msg="compute graph" device=CUDA0 size="128.0 MiB"
time=2025-12-20T05:21:41.916+02:00 level=INFO source=device.go:272 msg="total memory" size="535.4 MiB"
time=2025-12-20T05:21:41.951+02:00 level=INFO source=runner.go:965 msg="starting go runner"
load_backend: loaded CPU backend from C:\Users\User\AppData\Local\Programs\Ollama\lib\ollama\ggml-cpu-icelake.dll
ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 CUDA devices:
  Device 0: NVIDIA GeForce GTX 1650, compute capability 7.5, VMM: yes, ID: GPU-fb36d98c-caf3-51fa-1c88-c1a9ec3cf2e7
The following devices will have suboptimal performance due to a lack of tensor cores:
  Device 0: NVIDIA GeForce GTX 1650
Consider compiling with CMAKE_CUDA_ARCHITECTURES=61-virtual;80-virtual and DGGML_CUDA_FORCE_MMQ to force the use of the Pascal code for Turing.
load_backend: loaded CUDA backend from C:\Users\User\AppData\Local\Programs\Ollama\lib\ollama\cuda_v13\ggml-cuda.dll
time=2025-12-20T05:21:42.049+02:00 level=INFO source=ggml.go:104 msg=system CPU.0.SSE3=1 CPU.0.SSSE3=1 CPU.0.AVX=1 CPU.0.AVX2=1 CPU.0.F16C=1 CPU.0.FMA=1 CPU.0.BMI2=1 CPU.0.AVX512=1 CPU.0.AVX512_VBMI=1 CPU.0.AVX512_VNNI=1 CPU.0.LLAMAFILE=1 CPU.1.LLAMAFILE=1 CUDA.0.ARCHS=750,800,860,870,890,900,1000,1030,1100,1200,1210 CUDA.0.USE_GRAPHS=1 CUDA.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(clang)
time=2025-12-20T05:21:42.050+02:00 level=INFO source=runner.go:1001 msg="Server listening on 127.0.0.1:53712"
time=2025-12-20T05:21:42.058+02:00 level=INFO source=runner.go:895 msg=load request="{Operation:commit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Auto KvSize:4096 KvCacheType: NumThreads:4 GPULayers:17[ID:GPU-fb36d98c-caf3-51fa-1c88-c1a9ec3cf2e7 Layers:17(0..16)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
time=2025-12-20T05:21:42.058+02:00 level=INFO source=server.go:1338 msg="waiting for llama runner to start responding"
time=2025-12-20T05:21:42.058+02:00 level=INFO source=server.go:1372 msg="waiting for server to become available" status="llm server loading model"
ggml_backend_cuda_device_get_memory device GPU-fb36d98c-caf3-51fa-1c88-c1a9ec3cf2e7 utilizing NVML memory reporting free: 3909283840 total: 4294967296
llama_model_load_from_file_impl: using device CUDA0 (NVIDIA GeForce GTX 1650) (0000:01:00.0) - 3728 MiB free
llama_model_loader: loaded meta data with 34 key-value pairs and 148 tensors from C:\Users\User\.ollama\models\blobs\sha256-b7bfeab6495a1ae3ae78811c1297df9f301b35261ff9580d42fb30dc4dc9034b (version GGUF V3 (latest))
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv   0:                       general.architecture str              = lfm2
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = LFM2 350M
llama_model_loader: - kv   3:                           general.basename str              = LFM2
llama_model_loader: - kv   4:                         general.size_label str              = 350M
llama_model_loader: - kv   5:                            general.license str              = other
llama_model_loader: - kv   6:                       general.license.name str              = lfm1.0
llama_model_loader: - kv   7:                       general.license.link str              = LICENSE
llama_model_loader: - kv   8:                               general.tags arr[str,4]       = ["liquid", "lfm2", "edge", "text-gene...
llama_model_loader: - kv   9:                          general.languages arr[str,8]       = ["en", "ar", "zh", "fr", "de", "ja", ...
llama_model_loader: - kv  10:                           lfm2.block_count u32              = 16
llama_model_loader: - kv  11:                        lfm2.context_length u32              = 128000
llama_model_loader: - kv  12:                      lfm2.embedding_length u32              = 1024
llama_model_loader: - kv  13:                   lfm2.feed_forward_length u32              = 4608
llama_model_loader: - kv  14:                  lfm2.attention.head_count u32              = 16
llama_model_loader: - kv  15:               lfm2.attention.head_count_kv arr[i32,16]      = [0, 0, 8, 0, 0, 8, 0, 0, 8, 0, 8, 0, ...
llama_model_loader: - kv  16:                        lfm2.rope.freq_base f32              = 1000000.000000
llama_model_loader: - kv  17:                            lfm2.vocab_size u32              = 65536
llama_model_loader: - kv  18:                     lfm2.shortconv.l_cache u32              = 3
llama_model_loader: - kv  19:      lfm2.attention.layer_norm_rms_epsilon f32              = 0.000010
llama_model_loader: - kv  20:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  21:                         tokenizer.ggml.pre str              = lfm2
llama_model_loader: - kv  22:                      tokenizer.ggml.tokens arr[str,65536]   = ["<|pad|>", "<|startoftext|>", "<|end...
llama_model_loader: - kv  23:                  tokenizer.ggml.token_type arr[i32,65536]   = [3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, ...
llama_model_loader: - kv  24:                      tokenizer.ggml.merges arr[str,63683]   = ["Ċ Ċ", "Ċ ĊĊ", "ĊĊ Ċ", "Ċ Į..
llama_model_loader: - kv  25:                tokenizer.ggml.bos_token_id u32              = 1
llama_model_loader: - kv  26:                tokenizer.ggml.eos_token_id u32              = 7
llama_model_loader: - kv  27:            tokenizer.ggml.padding_token_id u32              = 0
llama_model_loader: - kv  28:               tokenizer.ggml.add_bos_token bool             = true
llama_model_loader: - kv  29:               tokenizer.ggml.add_sep_token bool             = false
llama_model_loader: - kv  30:               tokenizer.ggml.add_eos_token bool             = false
llama_model_loader: - kv  31:                    tokenizer.chat_template str              = {{bos_token}}{% for message in messag...
llama_model_loader: - kv  32:               general.quantization_version u32              = 2
llama_model_loader: - kv  33:                          general.file_type u32              = 7
llama_model_loader: - type  f32:   55 tensors
llama_model_loader: - type q8_0:   93 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type   = Q8_0
print_info: file size   = 359.37 MiB (8.50 BPW) 
load: printing all EOG tokens:
load:   - 2 ('<|endoftext|>')
load:   - 7 ('<|im_end|>')
load: special tokens cache size = 507
load: token to piece cache size = 0.3756 MB
print_info: arch             = lfm2
print_info: vocab_only       = 0
print_info: no_alloc         = 0
print_info: n_ctx_train      = 128000
print_info: n_embd           = 1024
print_info: n_embd_inp       = 1024
print_info: n_layer          = 16
print_info: n_head           = 16
print_info: n_head_kv        = [0, 0, 8, 0, 0, 8, 0, 0, 8, 0, 8, 0, 8, 0, 8, 0]
print_info: n_rot            = 64
print_info: n_swa            = 0
print_info: is_swa_any       = 0
print_info: n_embd_head_k    = 64
print_info: n_embd_head_v    = 64
print_info: n_gqa            = [0, 0, 2, 0, 0, 2, 0, 0, 2, 0, 2, 0, 2, 0, 2, 0]
print_info: n_embd_k_gqa     = [0, 0, 512, 0, 0, 512, 0, 0, 512, 0, 512, 0, 512, 0, 512, 0]
print_info: n_embd_v_gqa     = [0, 0, 512, 0, 0, 512, 0, 0, 512, 0, 512, 0, 512, 0, 512, 0]
print_info: f_norm_eps       = 0.0e+00
print_info: f_norm_rms_eps   = 1.0e-05
print_info: f_clamp_kqv      = 0.0e+00
print_info: f_max_alibi_bias = 0.0e+00
print_info: f_logit_scale    = 0.0e+00
print_info: f_attn_scale     = 0.0e+00
print_info: n_ff             = 4608
print_info: n_expert         = 0
print_info: n_expert_used    = 0
print_info: n_expert_groups  = 0
print_info: n_group_used     = 0
print_info: causal attn      = 1
print_info: pooling type     = 0
print_info: rope type        = 2
print_info: rope scaling     = linear
print_info: freq_base_train  = 1000000.0
print_info: freq_scale_train = 1
print_info: n_ctx_orig_yarn  = 128000
print_info: rope_yarn_log_mul= 0.0000
print_info: rope_finetuned   = unknown
print_info: model type       = 350M
print_info: model params     = 354.48 M
print_info: general.name     = LFM2 350M
print_info: vocab type       = BPE
print_info: n_vocab          = 65536
print_info: n_merges         = 63683
print_info: BOS token        = 1 '<|startoftext|>'
print_info: EOS token        = 7 '<|im_end|>'
print_info: EOT token        = 2 '<|endoftext|>'
print_info: PAD token        = 0 '<|pad|>'
print_info: LF token         = 708 'Ċ'
print_info: EOG token        = 2 '<|endoftext|>'
print_info: EOG token        = 7 '<|im_end|>'
print_info: max token length = 30
load_tensors: loading model tensors, this can take a while... (mmap = false)
llama_model_load: error loading model: missing tensor 'output_norm'
llama_model_load_from_file_impl: failed to load model
panic: unable to load model: C:\Users\User\.ollama\models\blobs\sha256-b7bfeab6495a1ae3ae78811c1297df9f301b35261ff9580d42fb30dc4dc9034b

goroutine 24 [running]:
github.com/ollama/ollama/runner/llamarunner.(*Server).loadModel(0xc000690280, {{0xc0002ede60, 0x1, 0x1}, 0x11, 0x0, 0x0, {0xc0002ede28, 0x1, 0x2}, ...}, ...)
	github.com/ollama/ollama/runner/llamarunner/runner.go:843 +0x33f
created by github.com/ollama/ollama/runner/llamarunner.(*Server).load in goroutine 51
	github.com/ollama/ollama/runner/llamarunner/runner.go:934 +0x889
time=2025-12-20T05:21:42.199+02:00 level=ERROR source=server.go:302 msg="llama runner terminated" error="exit status 2"
time=2025-12-20T05:21:42.309+02:00 level=INFO source=sched.go:470 msg="Load failed" model=C:\Users\User\.ollama\models\blobs\sha256-b7bfeab6495a1ae3ae78811c1297df9f301b35261ff9580d42fb30dc4dc9034b error="llama runner process has terminated: error loading model: missing tensor 'output_norm'"
[GIN] 2025/12/20 - 05:21:42 | 500 |    730.2481ms |       127.0.0.1 | POST     "/api/generate"

One notable weird thing I see is that it says cuda_v13 but my installed on system version is 12.8

<!-- gh-comment-id:3677324820 --> @qustrolabe commented on GitHub (Dec 20, 2025): I have the same issue with `sam860/LFM2:350m` and `sam860/LFM2:1.2b` models, trying some other models I have and they work flawlessly. I remember running those 2 broken ones around month ago without any problem before recent update. Win 11 ollama 0.13.5 Logs don't say much besides `error loading model: missing tensor 'output_norm'` <details> <summary>Logs</summary> ``` time=2025-12-20T05:21:40.813+02:00 level=INFO source=routes.go:1554 msg="server config" env="map[CUDA_VISIBLE_DEVICES: GGML_VK_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:4096 OLLAMA_DEBUG:INFO OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:C:\\Users\\User\\.ollama\\models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_REMOTES:[ollama.com] OLLAMA_SCHED_SPREAD:false OLLAMA_VULKAN:false ROCR_VISIBLE_DEVICES:]" time=2025-12-20T05:21:40.819+02:00 level=INFO source=images.go:493 msg="total blobs: 28" time=2025-12-20T05:21:40.821+02:00 level=INFO source=images.go:500 msg="total unused blobs removed: 0" time=2025-12-20T05:21:40.822+02:00 level=INFO source=routes.go:1607 msg="Listening on 127.0.0.1:11434 (version 0.13.5)" time=2025-12-20T05:21:40.823+02:00 level=INFO source=runner.go:67 msg="discovering available GPUs..." time=2025-12-20T05:21:40.840+02:00 level=INFO source=runner.go:106 msg="experimental Vulkan support disabled. To enable, set OLLAMA_VULKAN=1" time=2025-12-20T05:21:40.849+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 53677" time=2025-12-20T05:21:41.040+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 53682" time=2025-12-20T05:21:41.228+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 53689" time=2025-12-20T05:21:41.327+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 53695" time=2025-12-20T05:21:41.327+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 53696" time=2025-12-20T05:21:41.535+02:00 level=INFO source=types.go:42 msg="inference compute" id=GPU-fb36d98c-caf3-51fa-1c88-c1a9ec3cf2e7 filter_id="" library=CUDA compute=7.5 name=CUDA0 description="NVIDIA GeForce GTX 1650" libdirs=ollama,cuda_v13 driver=13.0 pci_id=0000:01:00.0 type=discrete total="4.0 GiB" available="3.6 GiB" time=2025-12-20T05:21:41.535+02:00 level=INFO source=routes.go:1648 msg="entering low vram mode" "total vram"="4.0 GiB" threshold="20.0 GiB" [GIN] 2025/12/20 - 05:21:41 | 200 | 0s | 127.0.0.1 | GET "/api/version" [GIN] 2025/12/20 - 05:21:41 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/12/20 - 05:21:41 | 200 | 24.5469ms | 127.0.0.1 | POST "/api/show" [GIN] 2025/12/20 - 05:21:41 | 200 | 24.4573ms | 127.0.0.1 | POST "/api/show" time=2025-12-20T05:21:41.623+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 53707" time=2025-12-20T05:21:41.814+02:00 level=INFO source=cpu_windows.go:148 msg=packages count=1 time=2025-12-20T05:21:41.814+02:00 level=INFO source=cpu_windows.go:195 msg="" package=0 cores=4 efficiency=0 threads=8 llama_model_loader: loaded meta data with 34 key-value pairs and 148 tensors from C:\Users\User\.ollama\models\blobs\sha256-b7bfeab6495a1ae3ae78811c1297df9f301b35261ff9580d42fb30dc4dc9034b (version GGUF V3 (latest)) llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. llama_model_loader: - kv 0: general.architecture str = lfm2 llama_model_loader: - kv 1: general.type str = model llama_model_loader: - kv 2: general.name str = LFM2 350M llama_model_loader: - kv 3: general.basename str = LFM2 llama_model_loader: - kv 4: general.size_label str = 350M llama_model_loader: - kv 5: general.license str = other llama_model_loader: - kv 6: general.license.name str = lfm1.0 llama_model_loader: - kv 7: general.license.link str = LICENSE llama_model_loader: - kv 8: general.tags arr[str,4] = ["liquid", "lfm2", "edge", "text-gene... llama_model_loader: - kv 9: general.languages arr[str,8] = ["en", "ar", "zh", "fr", "de", "ja", ... llama_model_loader: - kv 10: lfm2.block_count u32 = 16 llama_model_loader: - kv 11: lfm2.context_length u32 = 128000 llama_model_loader: - kv 12: lfm2.embedding_length u32 = 1024 llama_model_loader: - kv 13: lfm2.feed_forward_length u32 = 4608 llama_model_loader: - kv 14: lfm2.attention.head_count u32 = 16 llama_model_loader: - kv 15: lfm2.attention.head_count_kv arr[i32,16] = [0, 0, 8, 0, 0, 8, 0, 0, 8, 0, 8, 0, ... llama_model_loader: - kv 16: lfm2.rope.freq_base f32 = 1000000.000000 llama_model_loader: - kv 17: lfm2.vocab_size u32 = 65536 llama_model_loader: - kv 18: lfm2.shortconv.l_cache u32 = 3 llama_model_loader: - kv 19: lfm2.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 20: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 21: tokenizer.ggml.pre str = lfm2 llama_model_loader: - kv 22: tokenizer.ggml.tokens arr[str,65536] = ["<|pad|>", "<|startoftext|>", "<|end... llama_model_loader: - kv 23: tokenizer.ggml.token_type arr[i32,65536] = [3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, ... llama_model_loader: - kv 24: tokenizer.ggml.merges arr[str,63683] = ["Ċ Ċ", "Ċ ĊĊ", "ĊĊ Ċ", "Ċ Į.. llama_model_loader: - kv 25: tokenizer.ggml.bos_token_id u32 = 1 llama_model_loader: - kv 26: tokenizer.ggml.eos_token_id u32 = 7 llama_model_loader: - kv 27: tokenizer.ggml.padding_token_id u32 = 0 llama_model_loader: - kv 28: tokenizer.ggml.add_bos_token bool = true llama_model_loader: - kv 29: tokenizer.ggml.add_sep_token bool = false llama_model_loader: - kv 30: tokenizer.ggml.add_eos_token bool = false llama_model_loader: - kv 31: tokenizer.chat_template str = {{bos_token}}{% for message in messag... llama_model_loader: - kv 32: general.quantization_version u32 = 2 llama_model_loader: - kv 33: general.file_type u32 = 7 llama_model_loader: - type f32: 55 tensors llama_model_loader: - type q8_0: 93 tensors print_info: file format = GGUF V3 (latest) print_info: file type = Q8_0 print_info: file size = 359.37 MiB (8.50 BPW) load: printing all EOG tokens: load: - 2 ('<|endoftext|>') load: - 7 ('<|im_end|>') load: special tokens cache size = 507 load: token to piece cache size = 0.3756 MB print_info: arch = lfm2 print_info: vocab_only = 1 print_info: no_alloc = 0 print_info: model type = ?B print_info: model params = 354.48 M print_info: general.name = LFM2 350M print_info: vocab type = BPE print_info: n_vocab = 65536 print_info: n_merges = 63683 print_info: BOS token = 1 '<|startoftext|>' print_info: EOS token = 7 '<|im_end|>' print_info: EOT token = 2 '<|endoftext|>' print_info: PAD token = 0 '<|pad|>' print_info: LF token = 708 'Ċ' print_info: EOG token = 2 '<|endoftext|>' print_info: EOG token = 7 '<|im_end|>' print_info: max token length = 30 llama_model_load: vocab only - skipping tensors time=2025-12-20T05:21:41.912+02:00 level=INFO source=server.go:429 msg="starting runner" cmd="C:\\Users\\User\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --model C:\\Users\\User\\.ollama\\models\\blobs\\sha256-b7bfeab6495a1ae3ae78811c1297df9f301b35261ff9580d42fb30dc4dc9034b --port 53712" time=2025-12-20T05:21:41.916+02:00 level=INFO source=sched.go:443 msg="system memory" total="15.8 GiB" free="4.2 GiB" free_swap="6.8 GiB" time=2025-12-20T05:21:41.916+02:00 level=INFO source=sched.go:450 msg="gpu memory" id=GPU-fb36d98c-caf3-51fa-1c88-c1a9ec3cf2e7 library=CUDA available="3.2 GiB" free="3.6 GiB" minimum="457.0 MiB" overhead="0 B" time=2025-12-20T05:21:41.916+02:00 level=INFO source=server.go:496 msg="loading model" "model layers"=17 requested=-1 time=2025-12-20T05:21:41.916+02:00 level=INFO source=device.go:240 msg="model weights" device=CUDA0 size="359.4 MiB" time=2025-12-20T05:21:41.916+02:00 level=INFO source=device.go:251 msg="kv cache" device=CUDA0 size="48.0 MiB" time=2025-12-20T05:21:41.916+02:00 level=INFO source=device.go:262 msg="compute graph" device=CUDA0 size="128.0 MiB" time=2025-12-20T05:21:41.916+02:00 level=INFO source=device.go:272 msg="total memory" size="535.4 MiB" time=2025-12-20T05:21:41.951+02:00 level=INFO source=runner.go:965 msg="starting go runner" load_backend: loaded CPU backend from C:\Users\User\AppData\Local\Programs\Ollama\lib\ollama\ggml-cpu-icelake.dll ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no ggml_cuda_init: found 1 CUDA devices: Device 0: NVIDIA GeForce GTX 1650, compute capability 7.5, VMM: yes, ID: GPU-fb36d98c-caf3-51fa-1c88-c1a9ec3cf2e7 The following devices will have suboptimal performance due to a lack of tensor cores: Device 0: NVIDIA GeForce GTX 1650 Consider compiling with CMAKE_CUDA_ARCHITECTURES=61-virtual;80-virtual and DGGML_CUDA_FORCE_MMQ to force the use of the Pascal code for Turing. load_backend: loaded CUDA backend from C:\Users\User\AppData\Local\Programs\Ollama\lib\ollama\cuda_v13\ggml-cuda.dll time=2025-12-20T05:21:42.049+02:00 level=INFO source=ggml.go:104 msg=system CPU.0.SSE3=1 CPU.0.SSSE3=1 CPU.0.AVX=1 CPU.0.AVX2=1 CPU.0.F16C=1 CPU.0.FMA=1 CPU.0.BMI2=1 CPU.0.AVX512=1 CPU.0.AVX512_VBMI=1 CPU.0.AVX512_VNNI=1 CPU.0.LLAMAFILE=1 CPU.1.LLAMAFILE=1 CUDA.0.ARCHS=750,800,860,870,890,900,1000,1030,1100,1200,1210 CUDA.0.USE_GRAPHS=1 CUDA.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(clang) time=2025-12-20T05:21:42.050+02:00 level=INFO source=runner.go:1001 msg="Server listening on 127.0.0.1:53712" time=2025-12-20T05:21:42.058+02:00 level=INFO source=runner.go:895 msg=load request="{Operation:commit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Auto KvSize:4096 KvCacheType: NumThreads:4 GPULayers:17[ID:GPU-fb36d98c-caf3-51fa-1c88-c1a9ec3cf2e7 Layers:17(0..16)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" time=2025-12-20T05:21:42.058+02:00 level=INFO source=server.go:1338 msg="waiting for llama runner to start responding" time=2025-12-20T05:21:42.058+02:00 level=INFO source=server.go:1372 msg="waiting for server to become available" status="llm server loading model" ggml_backend_cuda_device_get_memory device GPU-fb36d98c-caf3-51fa-1c88-c1a9ec3cf2e7 utilizing NVML memory reporting free: 3909283840 total: 4294967296 llama_model_load_from_file_impl: using device CUDA0 (NVIDIA GeForce GTX 1650) (0000:01:00.0) - 3728 MiB free llama_model_loader: loaded meta data with 34 key-value pairs and 148 tensors from C:\Users\User\.ollama\models\blobs\sha256-b7bfeab6495a1ae3ae78811c1297df9f301b35261ff9580d42fb30dc4dc9034b (version GGUF V3 (latest)) llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. llama_model_loader: - kv 0: general.architecture str = lfm2 llama_model_loader: - kv 1: general.type str = model llama_model_loader: - kv 2: general.name str = LFM2 350M llama_model_loader: - kv 3: general.basename str = LFM2 llama_model_loader: - kv 4: general.size_label str = 350M llama_model_loader: - kv 5: general.license str = other llama_model_loader: - kv 6: general.license.name str = lfm1.0 llama_model_loader: - kv 7: general.license.link str = LICENSE llama_model_loader: - kv 8: general.tags arr[str,4] = ["liquid", "lfm2", "edge", "text-gene... llama_model_loader: - kv 9: general.languages arr[str,8] = ["en", "ar", "zh", "fr", "de", "ja", ... llama_model_loader: - kv 10: lfm2.block_count u32 = 16 llama_model_loader: - kv 11: lfm2.context_length u32 = 128000 llama_model_loader: - kv 12: lfm2.embedding_length u32 = 1024 llama_model_loader: - kv 13: lfm2.feed_forward_length u32 = 4608 llama_model_loader: - kv 14: lfm2.attention.head_count u32 = 16 llama_model_loader: - kv 15: lfm2.attention.head_count_kv arr[i32,16] = [0, 0, 8, 0, 0, 8, 0, 0, 8, 0, 8, 0, ... llama_model_loader: - kv 16: lfm2.rope.freq_base f32 = 1000000.000000 llama_model_loader: - kv 17: lfm2.vocab_size u32 = 65536 llama_model_loader: - kv 18: lfm2.shortconv.l_cache u32 = 3 llama_model_loader: - kv 19: lfm2.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 20: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 21: tokenizer.ggml.pre str = lfm2 llama_model_loader: - kv 22: tokenizer.ggml.tokens arr[str,65536] = ["<|pad|>", "<|startoftext|>", "<|end... llama_model_loader: - kv 23: tokenizer.ggml.token_type arr[i32,65536] = [3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, ... llama_model_loader: - kv 24: tokenizer.ggml.merges arr[str,63683] = ["Ċ Ċ", "Ċ ĊĊ", "ĊĊ Ċ", "Ċ Į.. llama_model_loader: - kv 25: tokenizer.ggml.bos_token_id u32 = 1 llama_model_loader: - kv 26: tokenizer.ggml.eos_token_id u32 = 7 llama_model_loader: - kv 27: tokenizer.ggml.padding_token_id u32 = 0 llama_model_loader: - kv 28: tokenizer.ggml.add_bos_token bool = true llama_model_loader: - kv 29: tokenizer.ggml.add_sep_token bool = false llama_model_loader: - kv 30: tokenizer.ggml.add_eos_token bool = false llama_model_loader: - kv 31: tokenizer.chat_template str = {{bos_token}}{% for message in messag... llama_model_loader: - kv 32: general.quantization_version u32 = 2 llama_model_loader: - kv 33: general.file_type u32 = 7 llama_model_loader: - type f32: 55 tensors llama_model_loader: - type q8_0: 93 tensors print_info: file format = GGUF V3 (latest) print_info: file type = Q8_0 print_info: file size = 359.37 MiB (8.50 BPW) load: printing all EOG tokens: load: - 2 ('<|endoftext|>') load: - 7 ('<|im_end|>') load: special tokens cache size = 507 load: token to piece cache size = 0.3756 MB print_info: arch = lfm2 print_info: vocab_only = 0 print_info: no_alloc = 0 print_info: n_ctx_train = 128000 print_info: n_embd = 1024 print_info: n_embd_inp = 1024 print_info: n_layer = 16 print_info: n_head = 16 print_info: n_head_kv = [0, 0, 8, 0, 0, 8, 0, 0, 8, 0, 8, 0, 8, 0, 8, 0] print_info: n_rot = 64 print_info: n_swa = 0 print_info: is_swa_any = 0 print_info: n_embd_head_k = 64 print_info: n_embd_head_v = 64 print_info: n_gqa = [0, 0, 2, 0, 0, 2, 0, 0, 2, 0, 2, 0, 2, 0, 2, 0] print_info: n_embd_k_gqa = [0, 0, 512, 0, 0, 512, 0, 0, 512, 0, 512, 0, 512, 0, 512, 0] print_info: n_embd_v_gqa = [0, 0, 512, 0, 0, 512, 0, 0, 512, 0, 512, 0, 512, 0, 512, 0] print_info: f_norm_eps = 0.0e+00 print_info: f_norm_rms_eps = 1.0e-05 print_info: f_clamp_kqv = 0.0e+00 print_info: f_max_alibi_bias = 0.0e+00 print_info: f_logit_scale = 0.0e+00 print_info: f_attn_scale = 0.0e+00 print_info: n_ff = 4608 print_info: n_expert = 0 print_info: n_expert_used = 0 print_info: n_expert_groups = 0 print_info: n_group_used = 0 print_info: causal attn = 1 print_info: pooling type = 0 print_info: rope type = 2 print_info: rope scaling = linear print_info: freq_base_train = 1000000.0 print_info: freq_scale_train = 1 print_info: n_ctx_orig_yarn = 128000 print_info: rope_yarn_log_mul= 0.0000 print_info: rope_finetuned = unknown print_info: model type = 350M print_info: model params = 354.48 M print_info: general.name = LFM2 350M print_info: vocab type = BPE print_info: n_vocab = 65536 print_info: n_merges = 63683 print_info: BOS token = 1 '<|startoftext|>' print_info: EOS token = 7 '<|im_end|>' print_info: EOT token = 2 '<|endoftext|>' print_info: PAD token = 0 '<|pad|>' print_info: LF token = 708 'Ċ' print_info: EOG token = 2 '<|endoftext|>' print_info: EOG token = 7 '<|im_end|>' print_info: max token length = 30 load_tensors: loading model tensors, this can take a while... (mmap = false) llama_model_load: error loading model: missing tensor 'output_norm' llama_model_load_from_file_impl: failed to load model panic: unable to load model: C:\Users\User\.ollama\models\blobs\sha256-b7bfeab6495a1ae3ae78811c1297df9f301b35261ff9580d42fb30dc4dc9034b goroutine 24 [running]: github.com/ollama/ollama/runner/llamarunner.(*Server).loadModel(0xc000690280, {{0xc0002ede60, 0x1, 0x1}, 0x11, 0x0, 0x0, {0xc0002ede28, 0x1, 0x2}, ...}, ...) github.com/ollama/ollama/runner/llamarunner/runner.go:843 +0x33f created by github.com/ollama/ollama/runner/llamarunner.(*Server).load in goroutine 51 github.com/ollama/ollama/runner/llamarunner/runner.go:934 +0x889 time=2025-12-20T05:21:42.199+02:00 level=ERROR source=server.go:302 msg="llama runner terminated" error="exit status 2" time=2025-12-20T05:21:42.309+02:00 level=INFO source=sched.go:470 msg="Load failed" model=C:\Users\User\.ollama\models\blobs\sha256-b7bfeab6495a1ae3ae78811c1297df9f301b35261ff9580d42fb30dc4dc9034b error="llama runner process has terminated: error loading model: missing tensor 'output_norm'" [GIN] 2025/12/20 - 05:21:42 | 500 | 730.2481ms | 127.0.0.1 | POST "/api/generate" ``` </details> One notable weird thing I see is that it says cuda_v13 but my installed on system version is 12.8
GiteaMirror commented 2026-04-29 09:09:43 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Dec 21, 2025):

https://github.com/ggml-org/llama.cpp/pull/18105

<!-- gh-comment-id:3678237553 --> @rick-github commented on GitHub (Dec 21, 2025): https://github.com/ggml-org/llama.cpp/pull/18105
GiteaMirror commented 2026-04-29 09:09:44 -05:00
Author
Owner
Copy Link

@lucas-vitrus commented on GitHub (Dec 23, 2025):

Is there a fix for this issue? I'm having the same problem

<!-- gh-comment-id:3685130105 --> @lucas-vitrus commented on GitHub (Dec 23, 2025): Is there a fix for this issue? I'm having the same problem
GiteaMirror commented 2026-04-29 09:09:45 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Dec 23, 2025):

Next vendor sync will fix it. In the meantime, either roll back to 0.13.4 or clone the ollama repo, patch in https://github.com/ggml-org/llama.cpp/pull/18105 and build a local version of ollama.

<!-- gh-comment-id:3685301591 --> @rick-github commented on GitHub (Dec 23, 2025): Next vendor sync will fix it. In the meantime, either roll back to 0.13.4 or clone the ollama repo, patch in https://github.com/ggml-org/llama.cpp/pull/18105 and [build](https://github.com/ollama/ollama/blob/main/docs/development.md) a local version of ollama.
GiteaMirror commented 2026-04-29 09:09:46 -05:00
Author
Owner
Copy Link

@rabol commented on GitHub (Dec 28, 2025):

sorry for asking, but when is next vendor sync?

<!-- gh-comment-id:3694461795 --> @rabol commented on GitHub (Dec 28, 2025): sorry for asking, but when is next vendor sync?
GiteaMirror commented 2026-04-29 09:09:47 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Dec 28, 2025):

#13570

<!-- gh-comment-id:3694607623 --> @rick-github commented on GitHub (Dec 28, 2025): #13570
GiteaMirror commented 2026-04-29 09:09:47 -05:00
Author
Owner
Copy Link

@jucleo-alexander commented on GitHub (Dec 31, 2025):

ollama version is 0.13.5
AMD Ryzen 7 5700G
62.09 GiB

journalctl -u ollama --no-pager --follow --pager-end

dic 31 11:40:59 t1 ollama[492660]: panic: unable to load model: /var/lib/ollama/.ollama/models/blobs/sha256-1f1d46904e25f1b67b538bd658ee4e11ed311864e5e8247b22ea5ab7488c83ee
dic 31 11:40:59 t1 ollama[492660]: goroutine 40 [running]:
dic 31 11:40:59 t1 ollama[492660]: github.com/ollama/ollama/runner/llamarunner.(*Server).loadModel(0xc000488640, {{0x0, 0x0, 0x0}, 0x0, 0x0, 0x0, {0x0, 0x0, 0x0}, ...}, ...)
dic 31 11:40:59 t1 ollama[492660]:         github.com/ollama/ollama/runner/llamarunner/runner.go:843 +0x33f
dic 31 11:40:59 t1 ollama[492660]: created by github.com/ollama/ollama/runner/llamarunner.(*Server).load in goroutine 12
dic 31 11:40:59 t1 ollama[492660]:         github.com/ollama/ollama/runner/llamarunner/runner.go:934 +0x889
dic 31 11:40:59 t1 ollama[492660]: time=2025-12-31T11:40:59.141-03:00 level=ERROR source=server.go:302 msg="llama runner terminated" error="exit status 2"
dic 31 11:40:59 t1 ollama[492660]: time=2025-12-31T11:40:59.306-03:00 level=INFO source=sched.go:470 msg="Load failed" model=/var/lib/ollama/.ollama/models/blobs/sha256-1f1d46904e25f1b67b538bd658ee4e11ed311864e5e8247b22ea5ab7488c83ee error="llama runner process has terminated: error loading model: missing tensor 'output_norm'\nllama_model_load_from_file_impl: failed to load model"
dic 31 11:40:59 t1 ollama[492660]: [GIN] 2025/12/31 - 11:40:59 | 500 |  411.924239ms |       127.0.0.1 | POST     "/api/generate"
dic 31 11:43:13 t1 ollama[492660]: [GIN] 2025/12/31 - 11:43:13 | 200 |      33.595µs |       127.0.0.1 | GET      "/api/version"
<!-- gh-comment-id:3702331349 --> @jucleo-alexander commented on GitHub (Dec 31, 2025): ollama version is 0.13.5 AMD Ryzen 7 5700G 62.09 GiB ```bash journalctl -u ollama --no-pager --follow --pager-end ``` ``` dic 31 11:40:59 t1 ollama[492660]: panic: unable to load model: /var/lib/ollama/.ollama/models/blobs/sha256-1f1d46904e25f1b67b538bd658ee4e11ed311864e5e8247b22ea5ab7488c83ee dic 31 11:40:59 t1 ollama[492660]: goroutine 40 [running]: dic 31 11:40:59 t1 ollama[492660]: github.com/ollama/ollama/runner/llamarunner.(*Server).loadModel(0xc000488640, {{0x0, 0x0, 0x0}, 0x0, 0x0, 0x0, {0x0, 0x0, 0x0}, ...}, ...) dic 31 11:40:59 t1 ollama[492660]: github.com/ollama/ollama/runner/llamarunner/runner.go:843 +0x33f dic 31 11:40:59 t1 ollama[492660]: created by github.com/ollama/ollama/runner/llamarunner.(*Server).load in goroutine 12 dic 31 11:40:59 t1 ollama[492660]: github.com/ollama/ollama/runner/llamarunner/runner.go:934 +0x889 dic 31 11:40:59 t1 ollama[492660]: time=2025-12-31T11:40:59.141-03:00 level=ERROR source=server.go:302 msg="llama runner terminated" error="exit status 2" dic 31 11:40:59 t1 ollama[492660]: time=2025-12-31T11:40:59.306-03:00 level=INFO source=sched.go:470 msg="Load failed" model=/var/lib/ollama/.ollama/models/blobs/sha256-1f1d46904e25f1b67b538bd658ee4e11ed311864e5e8247b22ea5ab7488c83ee error="llama runner process has terminated: error loading model: missing tensor 'output_norm'\nllama_model_load_from_file_impl: failed to load model" dic 31 11:40:59 t1 ollama[492660]: [GIN] 2025/12/31 - 11:40:59 | 500 | 411.924239ms | 127.0.0.1 | POST "/api/generate" dic 31 11:43:13 t1 ollama[492660]: [GIN] 2025/12/31 - 11:43:13 | 200 | 33.595µs | 127.0.0.1 | GET "/api/version" ```
GiteaMirror commented 2026-04-29 09:09:48 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Dec 31, 2025):

@jucleo-alexander https://github.com/ollama/ollama/issues/13529#issuecomment-3685301591

<!-- gh-comment-id:3702370253 --> @rick-github commented on GitHub (Dec 31, 2025): @jucleo-alexander https://github.com/ollama/ollama/issues/13529#issuecomment-3685301591
GiteaMirror commented 2026-04-29 09:09:51 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 14, 2026):

https://github.com/ollama/ollama/issues/13529#issuecomment-3685301591

<!-- gh-comment-id:3750659595 --> @rick-github commented on GitHub (Jan 14, 2026): https://github.com/ollama/ollama/issues/13529#issuecomment-3685301591
GiteaMirror commented 2026-04-29 09:09:53 -05:00
Author
Owner
Copy Link

@nise commented on GitHub (Feb 26, 2026):

#13529 was not merged in favour of #13832.
However, the error still comes up with any of the LFM models, e.g.

ollama --version
ollama version is 0.17.0
ollama run hf.co/LiquidAI/LFM2.5-Audio-1.5B-GGUF:F16
pulling manifest 
pulling 60c8b3c36e52: 100% ▕██████████████████▏ 2.3 GB                         
pulling 5188f2b355da: 100% ▕██████████████████▏  10 KB                         
pulling a776233427f3: 100% ▕██████████████████▏  197 B                         
pulling 71330d782076: 100% ▕██████████████████▏ 458 MB                         
pulling d7ca051a8aac: 100% ▕██████████████████▏   75 B                         
pulling 30d196b48087: 100% ▕██████████████████▏  771 B                         
verifying sha256 digest 
writing manifest 
success 
Error: 500 Internal Server Error: llama runner process has terminated: error loading model: missing tensor 'output_norm'
<!-- gh-comment-id:3965965672 --> @nise commented on GitHub (Feb 26, 2026): #13529 was not merged in favour of #13832. However, the error still comes up with any of the LFM models, e.g. ``` ollama --version ollama version is 0.17.0 ollama run hf.co/LiquidAI/LFM2.5-Audio-1.5B-GGUF:F16 pulling manifest pulling 60c8b3c36e52: 100% ▕██████████████████▏ 2.3 GB pulling 5188f2b355da: 100% ▕██████████████████▏ 10 KB pulling a776233427f3: 100% ▕██████████████████▏ 197 B pulling 71330d782076: 100% ▕██████████████████▏ 458 MB pulling d7ca051a8aac: 100% ▕██████████████████▏ 75 B pulling 30d196b48087: 100% ▕██████████████████▏ 771 B verifying sha256 digest writing manifest success Error: 500 Internal Server Error: llama runner process has terminated: error loading model: missing tensor 'output_norm' ```
GiteaMirror commented 2026-04-29 09:09:55 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 26, 2026):

ollama doesn't support audio models.

<!-- gh-comment-id:3967406785 --> @rick-github commented on GitHub (Feb 26, 2026): ollama doesn't support audio models.
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#55424
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?