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

[PR #14604] fix(metal): resolve static_assert crash on Apple M5 #77037

New Issue
Open
opened 2026-05-05 09:45:39 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-05-05 09:45:39 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/14604
Author: @Lumysia
Created: 3/4/2026
Status: 🔄 Open

Base: main ← Head: fix-m5-metal-static-assert

📝 Commits (2)

  • f69fabe fix(metal): remove incompatible bf16/f16 kernels for Apple M5
  • 963bb23 fix(metal): handle missing bf16_f16 kernels gracefully and fix dylib fallback

📊 Changes

5 files changed (+27 additions, -12 deletions)

View changed files

📝 ml/backend/ggml/ggml/src/ggml-metal/ggml-metal-device.cpp (+12 -0)
📝 ml/backend/ggml/ggml/src/ggml-metal/ggml-metal-device.m (+3 -0)
📝 ml/backend/ggml/ggml/src/ggml-metal/ggml-metal-embed.metal (+0 -6)
📝 ml/backend/ggml/ggml/src/ggml-metal/ggml-metal-ops.cpp (+12 -0)
📝 ml/backend/ggml/ggml/src/ggml-metal/ggml-metal.metal (+0 -6)

📄 Description

Fixed #14432, also resolves #13460

The Issue

This PR addresses a critical Metal backend crash on Apple M5 MTLGPUFamilyApple10 devices. The bf16xf16 mixed-precision matrix kernels trigger a static_assert failure in the Metal Performance Primitives header because the M5 hardware requires strict type matching for cooperative tensor operations.

The Fix

To align with upstream llama.cpp logic and ensure stability across different build environments, this PR implements the following:

  • Cleaned up Metal Shaders: Removed the incompatible kernel_mul_mm_bf16_f16 and its indirect variant kernel_mul_mm_id_bf16_f16 templates from both source and embedded .metal files to prevent the strict M5 compiler from asserting.
  • Graph Fallback: Updated ggml_metal_device_supports_op in ggml-metal-device.m to explicitly return false for BF16 x F16 multiplications. This tells the graph scheduler to safely handle this via a cast node or CPU fallback, rather than forcing it onto the Metal backend.
  • Dylib Fix: Added null-checks during pipeline initialization in ggml-metal-ops.cpp and ggml-metal-device.cpp. This prevents silent GPU discovery failures and null pointer crashes when the project is compiled with GGML_METAL=ON and attempts to look up the removed kernels.

Verification

  • Successfully tested on M5 running macOS 26.3 (25D125).
  • Tested models: llama3.2:3b and gemma3:270m-it-bf16.
  • Verified that models run perfectly on the GPU with this patch under both EMBED_LIBRARY=1 and EMBED_LIBRARY=OFF (dylib) build configurations without silent CPU fallbacks.

Logs

olama run llama3.2:3b

➜  ollama git:(fix-m5-metal-static-assert) ✗ ./ollama run llama3.2:3b
>>> hello
Hello! How can I assist you today?

>>> /bye

ollama serve

➜  ollama git:(fix-m5-metal-static-assert) ✗ ./ollama serve
time=2026-03-03T22:20:53.291-05:00 level=INFO source=routes.go:1743 msg="server config" env="map[HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:0 OLLAMA_DEBUG:INFO OLLAMA_EDITOR: OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 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:/Users/vya/.ollama/models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NO_CLOUD: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 http_proxy: https_proxy: no_proxy:]"
time=2026-03-03T22:20:53.291-05:00 level=INFO source=routes.go:1745 msg="Ollama cloud disabled: false"
time=2026-03-03T22:20:53.293-05:00 level=INFO source=images.go:477 msg="total blobs: 16"
time=2026-03-03T22:20:53.293-05:00 level=INFO source=images.go:484 msg="total unused blobs removed: 0"
[GIN-debug] [WARNING] Creating an Engine instance with the Logger and Recovery middleware already attached.

[GIN-debug] [WARNING] Running in "debug" mode. Switch to "release" mode in production.
 - using env:   export GIN_MODE=release
 - using code:  gin.SetMode(gin.ReleaseMode)

[GIN-debug] HEAD   /                         --> github.com/ollama/ollama/server.(*Server).GenerateRoutes.func1 (5 handlers)
[GIN-debug] GET    /                         --> github.com/ollama/ollama/server.(*Server).GenerateRoutes.func2 (5 handlers)
[GIN-debug] HEAD   /api/version              --> github.com/ollama/ollama/server.(*Server).GenerateRoutes.func3 (5 handlers)
[GIN-debug] GET    /api/version              --> github.com/ollama/ollama/server.(*Server).GenerateRoutes.func4 (5 handlers)
[GIN-debug] GET    /api/status               --> github.com/ollama/ollama/server.(*Server).StatusHandler-fm (5 handlers)
[GIN-debug] POST   /api/pull                 --> github.com/ollama/ollama/server.(*Server).PullHandler-fm (5 handlers)
[GIN-debug] POST   /api/push                 --> github.com/ollama/ollama/server.(*Server).PushHandler-fm (5 handlers)
[GIN-debug] HEAD   /api/tags                 --> github.com/ollama/ollama/server.(*Server).ListHandler-fm (5 handlers)
[GIN-debug] GET    /api/tags                 --> github.com/ollama/ollama/server.(*Server).ListHandler-fm (5 handlers)
[GIN-debug] POST   /api/show                 --> github.com/ollama/ollama/server.(*Server).ShowHandler-fm (5 handlers)
[GIN-debug] DELETE /api/delete               --> github.com/ollama/ollama/server.(*Server).DeleteHandler-fm (5 handlers)
[GIN-debug] POST   /api/me                   --> github.com/ollama/ollama/server.(*Server).WhoamiHandler-fm (5 handlers)
[GIN-debug] POST   /api/signout              --> github.com/ollama/ollama/server.(*Server).SignoutHandler-fm (5 handlers)
[GIN-debug] DELETE /api/user/keys/:encodedKey --> github.com/ollama/ollama/server.(*Server).SignoutHandler-fm (5 handlers)
[GIN-debug] POST   /api/create               --> github.com/ollama/ollama/server.(*Server).CreateHandler-fm (5 handlers)
[GIN-debug] POST   /api/blobs/:digest        --> github.com/ollama/ollama/server.(*Server).CreateBlobHandler-fm (5 handlers)
[GIN-debug] HEAD   /api/blobs/:digest        --> github.com/ollama/ollama/server.(*Server).HeadBlobHandler-fm (5 handlers)
[GIN-debug] POST   /api/copy                 --> github.com/ollama/ollama/server.(*Server).CopyHandler-fm (5 handlers)
[GIN-debug] GET    /api/experimental/aliases --> github.com/ollama/ollama/server.(*Server).ListAliasesHandler-fm (5 handlers)
[GIN-debug] POST   /api/experimental/aliases --> github.com/ollama/ollama/server.(*Server).CreateAliasHandler-fm (5 handlers)
[GIN-debug] DELETE /api/experimental/aliases --> github.com/ollama/ollama/server.(*Server).DeleteAliasHandler-fm (5 handlers)
[GIN-debug] GET    /api/ps                   --> github.com/ollama/ollama/server.(*Server).PsHandler-fm (5 handlers)
[GIN-debug] POST   /api/generate             --> github.com/ollama/ollama/server.(*Server).GenerateHandler-fm (5 handlers)
[GIN-debug] POST   /api/chat                 --> github.com/ollama/ollama/server.(*Server).ChatHandler-fm (5 handlers)
[GIN-debug] POST   /api/embed                --> github.com/ollama/ollama/server.(*Server).EmbedHandler-fm (5 handlers)
[GIN-debug] POST   /api/embeddings           --> github.com/ollama/ollama/server.(*Server).EmbeddingsHandler-fm (5 handlers)
[GIN-debug] POST   /v1/chat/completions      --> github.com/ollama/ollama/server.(*Server).ChatHandler-fm (7 handlers)
[GIN-debug] POST   /v1/completions           --> github.com/ollama/ollama/server.(*Server).GenerateHandler-fm (7 handlers)
[GIN-debug] POST   /v1/embeddings            --> github.com/ollama/ollama/server.(*Server).EmbedHandler-fm (7 handlers)
[GIN-debug] GET    /v1/models                --> github.com/ollama/ollama/server.(*Server).ListHandler-fm (6 handlers)
[GIN-debug] GET    /v1/models/:model         --> github.com/ollama/ollama/server.(*Server).ShowHandler-fm (7 handlers)
[GIN-debug] POST   /v1/responses             --> github.com/ollama/ollama/server.(*Server).ChatHandler-fm (7 handlers)
[GIN-debug] POST   /v1/images/generations    --> github.com/ollama/ollama/server.(*Server).GenerateHandler-fm (7 handlers)
[GIN-debug] POST   /v1/images/edits          --> github.com/ollama/ollama/server.(*Server).GenerateHandler-fm (7 handlers)
[GIN-debug] POST   /v1/messages              --> github.com/ollama/ollama/server.(*Server).ChatHandler-fm (7 handlers)
time=2026-03-03T22:20:53.294-05:00 level=INFO source=routes.go:1798 msg="Listening on [::]:11434 (version 0.0.0)"
time=2026-03-03T22:20:53.297-05:00 level=INFO source=runner.go:67 msg="discovering available GPUs..."
time=2026-03-03T22:20:53.299-05:00 level=INFO source=server.go:430 msg="starting runner" cmd="/Users/vya/Projects/ollama/ollama runner --ollama-engine --port 56232"
time=2026-03-03T22:21:01.210-05:00 level=INFO source=types.go:42 msg="inference compute" id=0 filter_id=0 library=Metal compute=0.0 name=Metal description="Apple M5" libdirs="" driver=0.0 pci_id="" type=discrete total="11.8 GiB" available="11.8 GiB"
time=2026-03-03T22:21:01.210-05:00 level=INFO source=routes.go:1848 msg="vram-based default context" total_vram="11.8 GiB" default_num_ctx=4096
[GIN] 2026/03/03 - 22:21:01 | 200 |      83.459µs |       127.0.0.1 | HEAD     "/"
[GIN] 2026/03/03 - 22:21:01 | 200 |   71.371083ms |       127.0.0.1 | POST     "/api/show"
[GIN] 2026/03/03 - 22:21:01 | 200 |      59.816ms |       127.0.0.1 | POST     "/api/show"
ggml_metal_device_init: testing tensor API for f16 support
ggml_metal_device_init: testing tensor API for bfloat support
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.006 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name:   Apple M5
ggml_metal_device_init: GPU family: MTLGPUFamilyApple10  (1010)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal4  (5002)
ggml_metal_device_init: simdgroup reduction   = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory    = true
ggml_metal_device_init: has bfloat            = true
ggml_metal_device_init: has tensor            = true
ggml_metal_device_init: use residency sets    = true
ggml_metal_device_init: use shared buffers    = true
ggml_metal_device_init: recommendedMaxWorkingSetSize  = 12713.12 MB
llama_model_load_from_file_impl: using device Metal (Apple M5) (unknown id) - 12123 MiB free
llama_model_loader: loaded meta data with 30 key-value pairs and 255 tensors from /Users/vya/.ollama/models/blobs/sha256-dde5aa3fc5ffc17176b5e8bdc82f587b24b2678c6c66101bf7da77af9f7ccdff (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              = llama
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = Llama 3.2 3B Instruct
llama_model_loader: - kv   3:                           general.finetune str              = Instruct
llama_model_loader: - kv   4:                           general.basename str              = Llama-3.2
llama_model_loader: - kv   5:                         general.size_label str              = 3B
llama_model_loader: - kv   6:                               general.tags arr[str,6]       = ["facebook", "meta", "pytorch", "llam...
llama_model_loader: - kv   7:                          general.languages arr[str,8]       = ["en", "de", "fr", "it", "pt", "hi", ...
llama_model_loader: - kv   8:                          llama.block_count u32              = 28
llama_model_loader: - kv   9:                       llama.context_length u32              = 131072
llama_model_loader: - kv  10:                     llama.embedding_length u32              = 3072
llama_model_loader: - kv  11:                  llama.feed_forward_length u32              = 8192
llama_model_loader: - kv  12:                 llama.attention.head_count u32              = 24
llama_model_loader: - kv  13:              llama.attention.head_count_kv u32              = 8
llama_model_loader: - kv  14:                       llama.rope.freq_base f32              = 500000.000000
llama_model_loader: - kv  15:     llama.attention.layer_norm_rms_epsilon f32              = 0.000010
llama_model_loader: - kv  16:                 llama.attention.key_length u32              = 128
llama_model_loader: - kv  17:               llama.attention.value_length u32              = 128
llama_model_loader: - kv  18:                          general.file_type u32              = 15
llama_model_loader: - kv  19:                           llama.vocab_size u32              = 128256
llama_model_loader: - kv  20:                 llama.rope.dimension_count u32              = 128
llama_model_loader: - kv  21:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  22:                         tokenizer.ggml.pre str              = llama-bpe
llama_model_loader: - kv  23:                      tokenizer.ggml.tokens arr[str,128256]  = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv  24:                  tokenizer.ggml.token_type arr[i32,128256]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv  25:                      tokenizer.ggml.merges arr[str,280147]  = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "...
llama_model_loader: - kv  26:                tokenizer.ggml.bos_token_id u32              = 128000
llama_model_loader: - kv  27:                tokenizer.ggml.eos_token_id u32              = 128009
llama_model_loader: - kv  28:                    tokenizer.chat_template str              = {{- bos_token }}\n{%- if custom_tools ...
llama_model_loader: - kv  29:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:   58 tensors
llama_model_loader: - type q4_K:  168 tensors
llama_model_loader: - type q6_K:   29 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type   = Q4_K - Medium
print_info: file size   = 1.87 GiB (5.01 BPW) 
load: printing all EOG tokens:
load:   - 128001 ('<|end_of_text|>')
load:   - 128008 ('<|eom_id|>')
load:   - 128009 ('<|eot_id|>')
load: special tokens cache size = 256
load: token to piece cache size = 0.7999 MB
print_info: arch             = llama
print_info: vocab_only       = 1
print_info: no_alloc         = 0
print_info: model type       = ?B
print_info: model params     = 3.21 B
print_info: general.name     = Llama 3.2 3B Instruct
print_info: vocab type       = BPE
print_info: n_vocab          = 128256
print_info: n_merges         = 280147
print_info: BOS token        = 128000 '<|begin_of_text|>'
print_info: EOS token        = 128009 '<|eot_id|>'
print_info: EOT token        = 128001 '<|end_of_text|>'
print_info: EOM token        = 128008 '<|eom_id|>'
print_info: LF token         = 198 'Ċ'
print_info: EOG token        = 128001 '<|end_of_text|>'
print_info: EOG token        = 128008 '<|eom_id|>'
print_info: EOG token        = 128009 '<|eot_id|>'
print_info: max token length = 256
llama_model_load: vocab only - skipping tensors
time=2026-03-03T22:21:01.620-05:00 level=INFO source=server.go:430 msg="starting runner" cmd="/Users/vya/Projects/ollama/ollama runner --model /Users/vya/.ollama/models/blobs/sha256-dde5aa3fc5ffc17176b5e8bdc82f587b24b2678c6c66101bf7da77af9f7ccdff --port 56239"
time=2026-03-03T22:21:01.622-05:00 level=INFO source=sched.go:489 msg="system memory" total="16.0 GiB" free="5.9 GiB" free_swap="0 B"
time=2026-03-03T22:21:01.623-05:00 level=INFO source=sched.go:496 msg="gpu memory" id=0 library=Metal available="11.3 GiB" free="11.8 GiB" minimum="512.0 MiB" overhead="0 B"
time=2026-03-03T22:21:01.623-05:00 level=INFO source=server.go:497 msg="loading model" "model layers"=29 requested=-1
time=2026-03-03T22:21:01.623-05:00 level=INFO source=device.go:240 msg="model weights" device=Metal size="1.9 GiB"
time=2026-03-03T22:21:01.623-05:00 level=INFO source=device.go:251 msg="kv cache" device=Metal size="448.0 MiB"
time=2026-03-03T22:21:01.623-05:00 level=INFO source=device.go:262 msg="compute graph" device=Metal size="256.5 MiB"
time=2026-03-03T22:21:01.623-05:00 level=INFO source=device.go:272 msg="total memory" size="2.6 GiB"
time=2026-03-03T22:21:01.634-05:00 level=INFO source=runner.go:965 msg="starting go runner"
ggml_metal_device_init: testing tensor API for f16 support
ggml_metal_device_init: testing tensor API for bfloat support
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 7.081 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name:   Apple M5
ggml_metal_device_init: GPU family: MTLGPUFamilyApple10  (1010)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal4  (5002)
ggml_metal_device_init: simdgroup reduction   = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory    = true
ggml_metal_device_init: has bfloat            = true
ggml_metal_device_init: has tensor            = true
ggml_metal_device_init: use residency sets    = true
ggml_metal_device_init: use shared buffers    = true
ggml_metal_device_init: recommendedMaxWorkingSetSize  = 12713.12 MB
time=2026-03-03T22:21:01.635-05:00 level=INFO source=ggml.go:104 msg=system Metal.0.EMBED_LIBRARY=1 CPU.0.NEON=1 CPU.0.ARM_FMA=1 CPU.0.FP16_VA=1 CPU.0.DOTPROD=1 CPU.0.LLAMAFILE=1 CPU.0.ACCELERATE=1 compiler=cgo(clang)
time=2026-03-03T22:21:08.750-05:00 level=INFO source=runner.go:1001 msg="Server listening on 127.0.0.1:56239"
time=2026-03-03T22:21:08.751-05:00 level=INFO source=runner.go:895 msg=load request="{Operation:commit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Auto KvSize:4096 KvCacheType: NumThreads:4 GPULayers:29[ID:0 Layers:29(0..28)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:true}"
llama_model_load_from_file_impl: using device Metal (Apple M5) (unknown id) - 12123 MiB free
time=2026-03-03T22:21:08.751-05:00 level=INFO source=server.go:1350 msg="waiting for llama runner to start responding"
time=2026-03-03T22:21:08.752-05:00 level=INFO source=server.go:1384 msg="waiting for server to become available" status="llm server loading model"
llama_model_loader: loaded meta data with 30 key-value pairs and 255 tensors from /Users/vya/.ollama/models/blobs/sha256-dde5aa3fc5ffc17176b5e8bdc82f587b24b2678c6c66101bf7da77af9f7ccdff (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              = llama
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = Llama 3.2 3B Instruct
llama_model_loader: - kv   3:                           general.finetune str              = Instruct
llama_model_loader: - kv   4:                           general.basename str              = Llama-3.2
llama_model_loader: - kv   5:                         general.size_label str              = 3B
llama_model_loader: - kv   6:                               general.tags arr[str,6]       = ["facebook", "meta", "pytorch", "llam...
llama_model_loader: - kv   7:                          general.languages arr[str,8]       = ["en", "de", "fr", "it", "pt", "hi", ...
llama_model_loader: - kv   8:                          llama.block_count u32              = 28
llama_model_loader: - kv   9:                       llama.context_length u32              = 131072
llama_model_loader: - kv  10:                     llama.embedding_length u32              = 3072
llama_model_loader: - kv  11:                  llama.feed_forward_length u32              = 8192
llama_model_loader: - kv  12:                 llama.attention.head_count u32              = 24
llama_model_loader: - kv  13:              llama.attention.head_count_kv u32              = 8
llama_model_loader: - kv  14:                       llama.rope.freq_base f32              = 500000.000000
llama_model_loader: - kv  15:     llama.attention.layer_norm_rms_epsilon f32              = 0.000010
llama_model_loader: - kv  16:                 llama.attention.key_length u32              = 128
llama_model_loader: - kv  17:               llama.attention.value_length u32              = 128
llama_model_loader: - kv  18:                          general.file_type u32              = 15
llama_model_loader: - kv  19:                           llama.vocab_size u32              = 128256
llama_model_loader: - kv  20:                 llama.rope.dimension_count u32              = 128
llama_model_loader: - kv  21:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  22:                         tokenizer.ggml.pre str              = llama-bpe
llama_model_loader: - kv  23:                      tokenizer.ggml.tokens arr[str,128256]  = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv  24:                  tokenizer.ggml.token_type arr[i32,128256]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv  25:                      tokenizer.ggml.merges arr[str,280147]  = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "...
llama_model_loader: - kv  26:                tokenizer.ggml.bos_token_id u32              = 128000
llama_model_loader: - kv  27:                tokenizer.ggml.eos_token_id u32              = 128009
llama_model_loader: - kv  28:                    tokenizer.chat_template str              = {{- bos_token }}\n{%- if custom_tools ...
llama_model_loader: - kv  29:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:   58 tensors
llama_model_loader: - type q4_K:  168 tensors
llama_model_loader: - type q6_K:   29 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type   = Q4_K - Medium
print_info: file size   = 1.87 GiB (5.01 BPW) 
load: printing all EOG tokens:
load:   - 128001 ('<|end_of_text|>')
load:   - 128008 ('<|eom_id|>')
load:   - 128009 ('<|eot_id|>')
load: special tokens cache size = 256
load: token to piece cache size = 0.7999 MB
print_info: arch             = llama
print_info: vocab_only       = 0
print_info: no_alloc         = 0
print_info: n_ctx_train      = 131072
print_info: n_embd           = 3072
print_info: n_embd_inp       = 3072
print_info: n_layer          = 28
print_info: n_head           = 24
print_info: n_head_kv        = 8
print_info: n_rot            = 128
print_info: n_swa            = 0
print_info: is_swa_any       = 0
print_info: n_embd_head_k    = 128
print_info: n_embd_head_v    = 128
print_info: n_gqa            = 3
print_info: n_embd_k_gqa     = 1024
print_info: n_embd_v_gqa     = 1024
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             = 8192
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        = 0
print_info: rope scaling     = linear
print_info: freq_base_train  = 500000.0
print_info: freq_scale_train = 1
print_info: n_ctx_orig_yarn  = 131072
print_info: rope_yarn_log_mul= 0.0000
print_info: rope_finetuned   = unknown
print_info: model type       = 3B
print_info: model params     = 3.21 B
print_info: general.name     = Llama 3.2 3B Instruct
print_info: vocab type       = BPE
print_info: n_vocab          = 128256
print_info: n_merges         = 280147
print_info: BOS token        = 128000 '<|begin_of_text|>'
print_info: EOS token        = 128009 '<|eot_id|>'
print_info: EOT token        = 128001 '<|end_of_text|>'
print_info: EOM token        = 128008 '<|eom_id|>'
print_info: LF token         = 198 'Ċ'
print_info: EOG token        = 128001 '<|end_of_text|>'
print_info: EOG token        = 128008 '<|eom_id|>'
print_info: EOG token        = 128009 '<|eot_id|>'
print_info: max token length = 256
load_tensors: loading model tensors, this can take a while... (mmap = true)
load_tensors: offloading 28 repeating layers to GPU
load_tensors: offloading output layer to GPU
load_tensors: offloaded 29/29 layers to GPU
load_tensors:   CPU_Mapped model buffer size =   308.23 MiB
load_tensors: Metal_Mapped model buffer size =  1918.35 MiB
llama_context: constructing llama_context
llama_context: n_seq_max     = 1
llama_context: n_ctx         = 4096
llama_context: n_ctx_seq     = 4096
llama_context: n_batch       = 512
llama_context: n_ubatch      = 512
llama_context: causal_attn   = 1
llama_context: flash_attn    = auto
llama_context: kv_unified    = false
llama_context: freq_base     = 500000.0
llama_context: freq_scale    = 1
llama_context: n_ctx_seq (4096) < n_ctx_train (131072) -- the full capacity of the model will not be utilized
ggml_metal_init: allocating
ggml_metal_init: picking default device: Apple M5
ggml_metal_init: use fusion         = true
ggml_metal_init: use concurrency    = true
ggml_metal_init: use graph optimize = true
llama_context:        CPU  output buffer size =     0.50 MiB
llama_kv_cache:      Metal KV buffer size =   448.00 MiB
llama_kv_cache: size =  448.00 MiB (  4096 cells,  28 layers,  1/1 seqs), K (f16):  224.00 MiB, V (f16):  224.00 MiB
llama_context: Flash Attention was auto, set to enabled
llama_context:      Metal compute buffer size =   256.50 MiB
llama_context:        CPU compute buffer size =    14.01 MiB
llama_context: graph nodes  = 875
llama_context: graph splits = 2
time=2026-03-03T22:21:10.007-05:00 level=INFO source=server.go:1388 msg="llama runner started in 8.38 seconds"
time=2026-03-03T22:21:10.007-05:00 level=INFO source=sched.go:565 msg="loaded runners" count=1
time=2026-03-03T22:21:10.007-05:00 level=INFO source=server.go:1350 msg="waiting for llama runner to start responding"
time=2026-03-03T22:21:10.007-05:00 level=INFO source=server.go:1388 msg="llama runner started in 8.38 seconds"
[GIN] 2026/03/03 - 22:21:10 | 200 |  8.662964167s |       127.0.0.1 | POST     "/api/generate"
[GIN] 2026/03/03 - 22:21:13 | 200 |  578.595292ms |       127.0.0.1 | POST     "/api/chat"

🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/ollama/ollama/pull/14604 **Author:** [@Lumysia](https://github.com/Lumysia) **Created:** 3/4/2026 **Status:** 🔄 Open **Base:** `main` ← **Head:** `fix-m5-metal-static-assert` --- ### 📝 Commits (2) - [`f69fabe`](https://github.com/ollama/ollama/commit/f69fabeecc19ac1150e2336a233169964e624519) fix(metal): remove incompatible bf16/f16 kernels for Apple M5 - [`963bb23`](https://github.com/ollama/ollama/commit/963bb23e603d1f3ff1716cf31211cd3f472f6d1c) fix(metal): handle missing bf16_f16 kernels gracefully and fix dylib fallback ### 📊 Changes **5 files changed** (+27 additions, -12 deletions) <details> <summary>View changed files</summary> 📝 `ml/backend/ggml/ggml/src/ggml-metal/ggml-metal-device.cpp` (+12 -0) 📝 `ml/backend/ggml/ggml/src/ggml-metal/ggml-metal-device.m` (+3 -0) 📝 `ml/backend/ggml/ggml/src/ggml-metal/ggml-metal-embed.metal` (+0 -6) 📝 `ml/backend/ggml/ggml/src/ggml-metal/ggml-metal-ops.cpp` (+12 -0) 📝 `ml/backend/ggml/ggml/src/ggml-metal/ggml-metal.metal` (+0 -6) </details> ### 📄 Description Fixed #14432, also resolves #13460 ## The Issue This PR addresses a critical Metal backend crash on Apple M5 MTLGPUFamilyApple10 devices. The bf16xf16 mixed-precision matrix kernels trigger a static_assert failure in the Metal Performance Primitives header because the M5 hardware requires strict type matching for cooperative tensor operations. ## The Fix To align with [upstream llama.cpp logic](https://github.com/ggml-org/llama.cpp/pull/18456) and ensure stability across different build environments, this PR implements the following: - Cleaned up Metal Shaders: Removed the incompatible kernel_mul_mm_bf16_f16 and its indirect variant kernel_mul_mm_id_bf16_f16 templates from both source and embedded .metal files to prevent the strict M5 compiler from asserting. - Graph Fallback: Updated ggml_metal_device_supports_op in ggml-metal-device.m to explicitly return false for BF16 x F16 multiplications. This tells the graph scheduler to safely handle this via a cast node or CPU fallback, rather than forcing it onto the Metal backend. - Dylib Fix: Added null-checks during pipeline initialization in ggml-metal-ops.cpp and ggml-metal-device.cpp. This prevents silent GPU discovery failures and null pointer crashes when the project is compiled with GGML_METAL=ON and attempts to look up the removed kernels. ## Verification - Successfully tested on M5 running macOS 26.3 (25D125). - Tested models: llama3.2:3b and gemma3:270m-it-bf16. - Verified that models run perfectly on the GPU with this patch under both EMBED_LIBRARY=1 and EMBED_LIBRARY=OFF (dylib) build configurations without silent CPU fallbacks. ## Logs > olama run llama3.2:3b ```bash ➜ ollama git:(fix-m5-metal-static-assert) ✗ ./ollama run llama3.2:3b >>> hello Hello! How can I assist you today? >>> /bye ``` > ollama serve ```bash ➜ ollama git:(fix-m5-metal-static-assert) ✗ ./ollama serve time=2026-03-03T22:20:53.291-05:00 level=INFO source=routes.go:1743 msg="server config" env="map[HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:0 OLLAMA_DEBUG:INFO OLLAMA_EDITOR: OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 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:/Users/vya/.ollama/models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NO_CLOUD: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 http_proxy: https_proxy: no_proxy:]" time=2026-03-03T22:20:53.291-05:00 level=INFO source=routes.go:1745 msg="Ollama cloud disabled: false" time=2026-03-03T22:20:53.293-05:00 level=INFO source=images.go:477 msg="total blobs: 16" time=2026-03-03T22:20:53.293-05:00 level=INFO source=images.go:484 msg="total unused blobs removed: 0" [GIN-debug] [WARNING] Creating an Engine instance with the Logger and Recovery middleware already attached. [GIN-debug] [WARNING] Running in "debug" mode. Switch to "release" mode in production. - using env: export GIN_MODE=release - using code: gin.SetMode(gin.ReleaseMode) [GIN-debug] HEAD / --> github.com/ollama/ollama/server.(*Server).GenerateRoutes.func1 (5 handlers) [GIN-debug] GET / --> github.com/ollama/ollama/server.(*Server).GenerateRoutes.func2 (5 handlers) [GIN-debug] HEAD /api/version --> github.com/ollama/ollama/server.(*Server).GenerateRoutes.func3 (5 handlers) [GIN-debug] GET /api/version --> github.com/ollama/ollama/server.(*Server).GenerateRoutes.func4 (5 handlers) [GIN-debug] GET /api/status --> github.com/ollama/ollama/server.(*Server).StatusHandler-fm (5 handlers) [GIN-debug] POST /api/pull --> github.com/ollama/ollama/server.(*Server).PullHandler-fm (5 handlers) [GIN-debug] POST /api/push --> github.com/ollama/ollama/server.(*Server).PushHandler-fm (5 handlers) [GIN-debug] HEAD /api/tags --> github.com/ollama/ollama/server.(*Server).ListHandler-fm (5 handlers) [GIN-debug] GET /api/tags --> github.com/ollama/ollama/server.(*Server).ListHandler-fm (5 handlers) [GIN-debug] POST /api/show --> github.com/ollama/ollama/server.(*Server).ShowHandler-fm (5 handlers) [GIN-debug] DELETE /api/delete --> github.com/ollama/ollama/server.(*Server).DeleteHandler-fm (5 handlers) [GIN-debug] POST /api/me --> github.com/ollama/ollama/server.(*Server).WhoamiHandler-fm (5 handlers) [GIN-debug] POST /api/signout --> github.com/ollama/ollama/server.(*Server).SignoutHandler-fm (5 handlers) [GIN-debug] DELETE /api/user/keys/:encodedKey --> github.com/ollama/ollama/server.(*Server).SignoutHandler-fm (5 handlers) [GIN-debug] POST /api/create --> github.com/ollama/ollama/server.(*Server).CreateHandler-fm (5 handlers) [GIN-debug] POST /api/blobs/:digest --> github.com/ollama/ollama/server.(*Server).CreateBlobHandler-fm (5 handlers) [GIN-debug] HEAD /api/blobs/:digest --> github.com/ollama/ollama/server.(*Server).HeadBlobHandler-fm (5 handlers) [GIN-debug] POST /api/copy --> github.com/ollama/ollama/server.(*Server).CopyHandler-fm (5 handlers) [GIN-debug] GET /api/experimental/aliases --> github.com/ollama/ollama/server.(*Server).ListAliasesHandler-fm (5 handlers) [GIN-debug] POST /api/experimental/aliases --> github.com/ollama/ollama/server.(*Server).CreateAliasHandler-fm (5 handlers) [GIN-debug] DELETE /api/experimental/aliases --> github.com/ollama/ollama/server.(*Server).DeleteAliasHandler-fm (5 handlers) [GIN-debug] GET /api/ps --> github.com/ollama/ollama/server.(*Server).PsHandler-fm (5 handlers) [GIN-debug] POST /api/generate --> github.com/ollama/ollama/server.(*Server).GenerateHandler-fm (5 handlers) [GIN-debug] POST /api/chat --> github.com/ollama/ollama/server.(*Server).ChatHandler-fm (5 handlers) [GIN-debug] POST /api/embed --> github.com/ollama/ollama/server.(*Server).EmbedHandler-fm (5 handlers) [GIN-debug] POST /api/embeddings --> github.com/ollama/ollama/server.(*Server).EmbeddingsHandler-fm (5 handlers) [GIN-debug] POST /v1/chat/completions --> github.com/ollama/ollama/server.(*Server).ChatHandler-fm (7 handlers) [GIN-debug] POST /v1/completions --> github.com/ollama/ollama/server.(*Server).GenerateHandler-fm (7 handlers) [GIN-debug] POST /v1/embeddings --> github.com/ollama/ollama/server.(*Server).EmbedHandler-fm (7 handlers) [GIN-debug] GET /v1/models --> github.com/ollama/ollama/server.(*Server).ListHandler-fm (6 handlers) [GIN-debug] GET /v1/models/:model --> github.com/ollama/ollama/server.(*Server).ShowHandler-fm (7 handlers) [GIN-debug] POST /v1/responses --> github.com/ollama/ollama/server.(*Server).ChatHandler-fm (7 handlers) [GIN-debug] POST /v1/images/generations --> github.com/ollama/ollama/server.(*Server).GenerateHandler-fm (7 handlers) [GIN-debug] POST /v1/images/edits --> github.com/ollama/ollama/server.(*Server).GenerateHandler-fm (7 handlers) [GIN-debug] POST /v1/messages --> github.com/ollama/ollama/server.(*Server).ChatHandler-fm (7 handlers) time=2026-03-03T22:20:53.294-05:00 level=INFO source=routes.go:1798 msg="Listening on [::]:11434 (version 0.0.0)" time=2026-03-03T22:20:53.297-05:00 level=INFO source=runner.go:67 msg="discovering available GPUs..." time=2026-03-03T22:20:53.299-05:00 level=INFO source=server.go:430 msg="starting runner" cmd="/Users/vya/Projects/ollama/ollama runner --ollama-engine --port 56232" time=2026-03-03T22:21:01.210-05:00 level=INFO source=types.go:42 msg="inference compute" id=0 filter_id=0 library=Metal compute=0.0 name=Metal description="Apple M5" libdirs="" driver=0.0 pci_id="" type=discrete total="11.8 GiB" available="11.8 GiB" time=2026-03-03T22:21:01.210-05:00 level=INFO source=routes.go:1848 msg="vram-based default context" total_vram="11.8 GiB" default_num_ctx=4096 [GIN] 2026/03/03 - 22:21:01 | 200 | 83.459µs | 127.0.0.1 | HEAD "/" [GIN] 2026/03/03 - 22:21:01 | 200 | 71.371083ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/03/03 - 22:21:01 | 200 | 59.816ms | 127.0.0.1 | POST "/api/show" ggml_metal_device_init: testing tensor API for f16 support ggml_metal_device_init: testing tensor API for bfloat support ggml_metal_library_init: using embedded metal library ggml_metal_library_init: loaded in 0.006 sec ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s) ggml_metal_device_init: GPU name: Apple M5 ggml_metal_device_init: GPU family: MTLGPUFamilyApple10 (1010) ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003) ggml_metal_device_init: GPU family: MTLGPUFamilyMetal4 (5002) ggml_metal_device_init: simdgroup reduction = true ggml_metal_device_init: simdgroup matrix mul. = true ggml_metal_device_init: has unified memory = true ggml_metal_device_init: has bfloat = true ggml_metal_device_init: has tensor = true ggml_metal_device_init: use residency sets = true ggml_metal_device_init: use shared buffers = true ggml_metal_device_init: recommendedMaxWorkingSetSize = 12713.12 MB llama_model_load_from_file_impl: using device Metal (Apple M5) (unknown id) - 12123 MiB free llama_model_loader: loaded meta data with 30 key-value pairs and 255 tensors from /Users/vya/.ollama/models/blobs/sha256-dde5aa3fc5ffc17176b5e8bdc82f587b24b2678c6c66101bf7da77af9f7ccdff (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 = llama llama_model_loader: - kv 1: general.type str = model llama_model_loader: - kv 2: general.name str = Llama 3.2 3B Instruct llama_model_loader: - kv 3: general.finetune str = Instruct llama_model_loader: - kv 4: general.basename str = Llama-3.2 llama_model_loader: - kv 5: general.size_label str = 3B llama_model_loader: - kv 6: general.tags arr[str,6] = ["facebook", "meta", "pytorch", "llam... llama_model_loader: - kv 7: general.languages arr[str,8] = ["en", "de", "fr", "it", "pt", "hi", ... llama_model_loader: - kv 8: llama.block_count u32 = 28 llama_model_loader: - kv 9: llama.context_length u32 = 131072 llama_model_loader: - kv 10: llama.embedding_length u32 = 3072 llama_model_loader: - kv 11: llama.feed_forward_length u32 = 8192 llama_model_loader: - kv 12: llama.attention.head_count u32 = 24 llama_model_loader: - kv 13: llama.attention.head_count_kv u32 = 8 llama_model_loader: - kv 14: llama.rope.freq_base f32 = 500000.000000 llama_model_loader: - kv 15: llama.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 16: llama.attention.key_length u32 = 128 llama_model_loader: - kv 17: llama.attention.value_length u32 = 128 llama_model_loader: - kv 18: general.file_type u32 = 15 llama_model_loader: - kv 19: llama.vocab_size u32 = 128256 llama_model_loader: - kv 20: llama.rope.dimension_count u32 = 128 llama_model_loader: - kv 21: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 22: tokenizer.ggml.pre str = llama-bpe llama_model_loader: - kv 23: tokenizer.ggml.tokens arr[str,128256] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 24: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 25: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "... llama_model_loader: - kv 26: tokenizer.ggml.bos_token_id u32 = 128000 llama_model_loader: - kv 27: tokenizer.ggml.eos_token_id u32 = 128009 llama_model_loader: - kv 28: tokenizer.chat_template str = {{- bos_token }}\n{%- if custom_tools ... llama_model_loader: - kv 29: general.quantization_version u32 = 2 llama_model_loader: - type f32: 58 tensors llama_model_loader: - type q4_K: 168 tensors llama_model_loader: - type q6_K: 29 tensors print_info: file format = GGUF V3 (latest) print_info: file type = Q4_K - Medium print_info: file size = 1.87 GiB (5.01 BPW) load: printing all EOG tokens: load: - 128001 ('<|end_of_text|>') load: - 128008 ('<|eom_id|>') load: - 128009 ('<|eot_id|>') load: special tokens cache size = 256 load: token to piece cache size = 0.7999 MB print_info: arch = llama print_info: vocab_only = 1 print_info: no_alloc = 0 print_info: model type = ?B print_info: model params = 3.21 B print_info: general.name = Llama 3.2 3B Instruct print_info: vocab type = BPE print_info: n_vocab = 128256 print_info: n_merges = 280147 print_info: BOS token = 128000 '<|begin_of_text|>' print_info: EOS token = 128009 '<|eot_id|>' print_info: EOT token = 128001 '<|end_of_text|>' print_info: EOM token = 128008 '<|eom_id|>' print_info: LF token = 198 'Ċ' print_info: EOG token = 128001 '<|end_of_text|>' print_info: EOG token = 128008 '<|eom_id|>' print_info: EOG token = 128009 '<|eot_id|>' print_info: max token length = 256 llama_model_load: vocab only - skipping tensors time=2026-03-03T22:21:01.620-05:00 level=INFO source=server.go:430 msg="starting runner" cmd="/Users/vya/Projects/ollama/ollama runner --model /Users/vya/.ollama/models/blobs/sha256-dde5aa3fc5ffc17176b5e8bdc82f587b24b2678c6c66101bf7da77af9f7ccdff --port 56239" time=2026-03-03T22:21:01.622-05:00 level=INFO source=sched.go:489 msg="system memory" total="16.0 GiB" free="5.9 GiB" free_swap="0 B" time=2026-03-03T22:21:01.623-05:00 level=INFO source=sched.go:496 msg="gpu memory" id=0 library=Metal available="11.3 GiB" free="11.8 GiB" minimum="512.0 MiB" overhead="0 B" time=2026-03-03T22:21:01.623-05:00 level=INFO source=server.go:497 msg="loading model" "model layers"=29 requested=-1 time=2026-03-03T22:21:01.623-05:00 level=INFO source=device.go:240 msg="model weights" device=Metal size="1.9 GiB" time=2026-03-03T22:21:01.623-05:00 level=INFO source=device.go:251 msg="kv cache" device=Metal size="448.0 MiB" time=2026-03-03T22:21:01.623-05:00 level=INFO source=device.go:262 msg="compute graph" device=Metal size="256.5 MiB" time=2026-03-03T22:21:01.623-05:00 level=INFO source=device.go:272 msg="total memory" size="2.6 GiB" time=2026-03-03T22:21:01.634-05:00 level=INFO source=runner.go:965 msg="starting go runner" ggml_metal_device_init: testing tensor API for f16 support ggml_metal_device_init: testing tensor API for bfloat support ggml_metal_library_init: using embedded metal library ggml_metal_library_init: loaded in 7.081 sec ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s) ggml_metal_device_init: GPU name: Apple M5 ggml_metal_device_init: GPU family: MTLGPUFamilyApple10 (1010) ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003) ggml_metal_device_init: GPU family: MTLGPUFamilyMetal4 (5002) ggml_metal_device_init: simdgroup reduction = true ggml_metal_device_init: simdgroup matrix mul. = true ggml_metal_device_init: has unified memory = true ggml_metal_device_init: has bfloat = true ggml_metal_device_init: has tensor = true ggml_metal_device_init: use residency sets = true ggml_metal_device_init: use shared buffers = true ggml_metal_device_init: recommendedMaxWorkingSetSize = 12713.12 MB time=2026-03-03T22:21:01.635-05:00 level=INFO source=ggml.go:104 msg=system Metal.0.EMBED_LIBRARY=1 CPU.0.NEON=1 CPU.0.ARM_FMA=1 CPU.0.FP16_VA=1 CPU.0.DOTPROD=1 CPU.0.LLAMAFILE=1 CPU.0.ACCELERATE=1 compiler=cgo(clang) time=2026-03-03T22:21:08.750-05:00 level=INFO source=runner.go:1001 msg="Server listening on 127.0.0.1:56239" time=2026-03-03T22:21:08.751-05:00 level=INFO source=runner.go:895 msg=load request="{Operation:commit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Auto KvSize:4096 KvCacheType: NumThreads:4 GPULayers:29[ID:0 Layers:29(0..28)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:true}" llama_model_load_from_file_impl: using device Metal (Apple M5) (unknown id) - 12123 MiB free time=2026-03-03T22:21:08.751-05:00 level=INFO source=server.go:1350 msg="waiting for llama runner to start responding" time=2026-03-03T22:21:08.752-05:00 level=INFO source=server.go:1384 msg="waiting for server to become available" status="llm server loading model" llama_model_loader: loaded meta data with 30 key-value pairs and 255 tensors from /Users/vya/.ollama/models/blobs/sha256-dde5aa3fc5ffc17176b5e8bdc82f587b24b2678c6c66101bf7da77af9f7ccdff (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 = llama llama_model_loader: - kv 1: general.type str = model llama_model_loader: - kv 2: general.name str = Llama 3.2 3B Instruct llama_model_loader: - kv 3: general.finetune str = Instruct llama_model_loader: - kv 4: general.basename str = Llama-3.2 llama_model_loader: - kv 5: general.size_label str = 3B llama_model_loader: - kv 6: general.tags arr[str,6] = ["facebook", "meta", "pytorch", "llam... llama_model_loader: - kv 7: general.languages arr[str,8] = ["en", "de", "fr", "it", "pt", "hi", ... llama_model_loader: - kv 8: llama.block_count u32 = 28 llama_model_loader: - kv 9: llama.context_length u32 = 131072 llama_model_loader: - kv 10: llama.embedding_length u32 = 3072 llama_model_loader: - kv 11: llama.feed_forward_length u32 = 8192 llama_model_loader: - kv 12: llama.attention.head_count u32 = 24 llama_model_loader: - kv 13: llama.attention.head_count_kv u32 = 8 llama_model_loader: - kv 14: llama.rope.freq_base f32 = 500000.000000 llama_model_loader: - kv 15: llama.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 16: llama.attention.key_length u32 = 128 llama_model_loader: - kv 17: llama.attention.value_length u32 = 128 llama_model_loader: - kv 18: general.file_type u32 = 15 llama_model_loader: - kv 19: llama.vocab_size u32 = 128256 llama_model_loader: - kv 20: llama.rope.dimension_count u32 = 128 llama_model_loader: - kv 21: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 22: tokenizer.ggml.pre str = llama-bpe llama_model_loader: - kv 23: tokenizer.ggml.tokens arr[str,128256] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 24: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 25: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "... llama_model_loader: - kv 26: tokenizer.ggml.bos_token_id u32 = 128000 llama_model_loader: - kv 27: tokenizer.ggml.eos_token_id u32 = 128009 llama_model_loader: - kv 28: tokenizer.chat_template str = {{- bos_token }}\n{%- if custom_tools ... llama_model_loader: - kv 29: general.quantization_version u32 = 2 llama_model_loader: - type f32: 58 tensors llama_model_loader: - type q4_K: 168 tensors llama_model_loader: - type q6_K: 29 tensors print_info: file format = GGUF V3 (latest) print_info: file type = Q4_K - Medium print_info: file size = 1.87 GiB (5.01 BPW) load: printing all EOG tokens: load: - 128001 ('<|end_of_text|>') load: - 128008 ('<|eom_id|>') load: - 128009 ('<|eot_id|>') load: special tokens cache size = 256 load: token to piece cache size = 0.7999 MB print_info: arch = llama print_info: vocab_only = 0 print_info: no_alloc = 0 print_info: n_ctx_train = 131072 print_info: n_embd = 3072 print_info: n_embd_inp = 3072 print_info: n_layer = 28 print_info: n_head = 24 print_info: n_head_kv = 8 print_info: n_rot = 128 print_info: n_swa = 0 print_info: is_swa_any = 0 print_info: n_embd_head_k = 128 print_info: n_embd_head_v = 128 print_info: n_gqa = 3 print_info: n_embd_k_gqa = 1024 print_info: n_embd_v_gqa = 1024 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 = 8192 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 = 0 print_info: rope scaling = linear print_info: freq_base_train = 500000.0 print_info: freq_scale_train = 1 print_info: n_ctx_orig_yarn = 131072 print_info: rope_yarn_log_mul= 0.0000 print_info: rope_finetuned = unknown print_info: model type = 3B print_info: model params = 3.21 B print_info: general.name = Llama 3.2 3B Instruct print_info: vocab type = BPE print_info: n_vocab = 128256 print_info: n_merges = 280147 print_info: BOS token = 128000 '<|begin_of_text|>' print_info: EOS token = 128009 '<|eot_id|>' print_info: EOT token = 128001 '<|end_of_text|>' print_info: EOM token = 128008 '<|eom_id|>' print_info: LF token = 198 'Ċ' print_info: EOG token = 128001 '<|end_of_text|>' print_info: EOG token = 128008 '<|eom_id|>' print_info: EOG token = 128009 '<|eot_id|>' print_info: max token length = 256 load_tensors: loading model tensors, this can take a while... (mmap = true) load_tensors: offloading 28 repeating layers to GPU load_tensors: offloading output layer to GPU load_tensors: offloaded 29/29 layers to GPU load_tensors: CPU_Mapped model buffer size = 308.23 MiB load_tensors: Metal_Mapped model buffer size = 1918.35 MiB llama_context: constructing llama_context llama_context: n_seq_max = 1 llama_context: n_ctx = 4096 llama_context: n_ctx_seq = 4096 llama_context: n_batch = 512 llama_context: n_ubatch = 512 llama_context: causal_attn = 1 llama_context: flash_attn = auto llama_context: kv_unified = false llama_context: freq_base = 500000.0 llama_context: freq_scale = 1 llama_context: n_ctx_seq (4096) < n_ctx_train (131072) -- the full capacity of the model will not be utilized ggml_metal_init: allocating ggml_metal_init: picking default device: Apple M5 ggml_metal_init: use fusion = true ggml_metal_init: use concurrency = true ggml_metal_init: use graph optimize = true llama_context: CPU output buffer size = 0.50 MiB llama_kv_cache: Metal KV buffer size = 448.00 MiB llama_kv_cache: size = 448.00 MiB ( 4096 cells, 28 layers, 1/1 seqs), K (f16): 224.00 MiB, V (f16): 224.00 MiB llama_context: Flash Attention was auto, set to enabled llama_context: Metal compute buffer size = 256.50 MiB llama_context: CPU compute buffer size = 14.01 MiB llama_context: graph nodes = 875 llama_context: graph splits = 2 time=2026-03-03T22:21:10.007-05:00 level=INFO source=server.go:1388 msg="llama runner started in 8.38 seconds" time=2026-03-03T22:21:10.007-05:00 level=INFO source=sched.go:565 msg="loaded runners" count=1 time=2026-03-03T22:21:10.007-05:00 level=INFO source=server.go:1350 msg="waiting for llama runner to start responding" time=2026-03-03T22:21:10.007-05:00 level=INFO source=server.go:1388 msg="llama runner started in 8.38 seconds" [GIN] 2026/03/03 - 22:21:10 | 200 | 8.662964167s | 127.0.0.1 | POST "/api/generate" [GIN] 2026/03/03 - 22:21:13 | 200 | 578.595292ms | 127.0.0.1 | POST "/api/chat" ``` --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
GiteaMirror added the pull-request label 2026-05-05 09:45:39 -05:00
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 pull-request
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#77037
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?