Eval bug: Memory allocation failure if any layer is on my second GPU (AMD official drivers) #15105
Description
Name and Version
version: 6096 (fd1234c)
built with clang version 19.1.5 for x86_64-pc-windows-msvc
Operating systems
Windows
GGML backends
Vulkan
Hardware
RTX 4060 Ti + RX 7900 XTX
Models
gpt-oss-120B
Problem description & steps to reproduce
Successful commands:
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 8 --main-gpu 1 --tensor-split 6,0
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 6 --main-gpu 1 --tensor-split 1,0
Failed commands:
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 4 --main-gpu 1
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 6 --main-gpu 0 --tensor-split 1,0
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 7 --main-gpu 0 --tensor-split 1,0
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 7 --main-gpu 1 --tensor-split 6,1
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 8 --main-gpu 1 --tensor-split 0,1
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 8 --main-gpu 1 --tensor-split 6,2
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 10 --main-gpu 1
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 11 --main-gpu 1 --tensor-split 0,1
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 16 --main-gpu 0 --tensor-split 4,12
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 16 --main-gpu 0 --tensor-split 6,10
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 16 --main-gpu 1
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 16 --main-gpu 1 --tensor-split 6,10
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 18
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 18 --main-gpu 1
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 18 --main-gpu 1 --tensor-split 6,12
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 20
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 21
vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 22
First Bad Commit
No response
Relevant log output
C:\AI>vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 11 --main-gpu 1 --tensor-split 0,1
load_backend: loaded RPC backend from C:\AI\vulkan\ggml-rpc.dll
ggml_vulkan: Found 2 Vulkan devices:
ggml_vulkan: 0 = NVIDIA GeForce RTX 4060 Ti (NVIDIA) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 49152 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 1 = AMD Radeon RX 7900 XTX (AMD proprietary driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
load_backend: loaded Vulkan backend from C:\AI\vulkan\ggml-vulkan.dll
load_backend: loaded CPU backend from C:\AI\vulkan\ggml-cpu-icelake.dll
build: 6096 (fd1234cb) with clang version 19.1.5 for x86_64-pc-windows-msvc
system info: n_threads = 6, n_threads_batch = 6, total_threads = 12
system_info: n_threads = 6 (n_threads_batch = 6) / 12 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | BMI2 = 1 | AVX512 = 1 | AVX512_VBMI = 1 | AVX512_VNNI = 1 | LLAMAFILE = 1 | OPENMP = 1 | REPACK = 1 |
main: binding port with default address family
main: HTTP server is listening, hostname: 127.0.0.1, port: 8080, http threads: 11
main: loading model
srv load_model: loading model 'gpt-oss-120b-F16.gguf'
llama_model_load_from_file_impl: using device Vulkan0 (NVIDIA GeForce RTX 4060 Ti) - 16109 MiB free
llama_model_load_from_file_impl: using device Vulkan1 (AMD Radeon RX 7900 XTX) - 24560 MiB free
llama_model_loader: loaded meta data with 41 key-value pairs and 687 tensors from gpt-oss-120b-F16.gguf (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 = gpt-oss
llama_model_loader: - kv 1: general.type str = model
llama_model_loader: - kv 2: general.name str = Gpt-Oss-120B
llama_model_loader: - kv 3: general.basename str = Gpt-Oss-120B
llama_model_loader: - kv 4: general.quantized_by str = Unsloth
llama_model_loader: - kv 5: general.size_label str = 120B
llama_model_loader: - kv 6: general.license str = apache-2.0
llama_model_loader: - kv 7: general.repo_url str = https://huggingface.co/unsloth
llama_model_loader: - kv 8: general.base_model.count u32 = 1
llama_model_loader: - kv 9: general.base_model.0.name str = Gpt Oss 120b
llama_model_loader: - kv 10: general.base_model.0.organization str = Openai
llama_model_loader: - kv 11: general.base_model.0.repo_url str = https://huggingface.co/openai/gpt-oss...
llama_model_loader: - kv 12: general.tags arr[str,3] = ["openai", "unsloth", "text-generation"]
llama_model_loader: - kv 13: gpt-oss.block_count u32 = 36
llama_model_loader: - kv 14: gpt-oss.context_length u32 = 131072
llama_model_loader: - kv 15: gpt-oss.embedding_length u32 = 2880
llama_model_loader: - kv 16: gpt-oss.feed_forward_length u32 = 2880
llama_model_loader: - kv 17: gpt-oss.attention.head_count u32 = 64
llama_model_loader: - kv 18: gpt-oss.attention.head_count_kv u32 = 8
llama_model_loader: - kv 19: gpt-oss.rope.freq_base f32 = 150000.000000
llama_model_loader: - kv 20: gpt-oss.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 21: gpt-oss.expert_count u32 = 128
llama_model_loader: - kv 22: gpt-oss.expert_used_count u32 = 4
llama_model_loader: - kv 23: gpt-oss.attention.key_length u32 = 64
llama_model_loader: - kv 24: gpt-oss.attention.value_length u32 = 64
llama_model_loader: - kv 25: general.file_type u32 = 1
llama_model_loader: - kv 26: gpt-oss.attention.sliding_window u32 = 128
llama_model_loader: - kv 27: gpt-oss.expert_feed_forward_length u32 = 2880
llama_model_loader: - kv 28: gpt-oss.rope.scaling.type str = yarn
llama_model_loader: - kv 29: gpt-oss.rope.scaling.factor f32 = 32.000000
llama_model_loader: - kv 30: gpt-oss.rope.scaling.original_context_length u32 = 4096
llama_model_loader: - kv 31: general.quantization_version u32 = 2
llama_model_loader: - kv 32: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 33: tokenizer.ggml.pre str = gpt-4o
llama_model_loader: - kv 34: tokenizer.ggml.tokens arr[str,201088] = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 35: tokenizer.ggml.token_type arr[i32,201088] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 36: tokenizer.ggml.merges arr[str,446189] = ["─á ─á", "─á ─á─á─á", "─á─á ─á─á", "...
llama_model_loader: - kv 37: tokenizer.ggml.bos_token_id u32 = 199998
llama_model_loader: - kv 38: tokenizer.ggml.eos_token_id u32 = 200002
llama_model_loader: - kv 39: tokenizer.ggml.padding_token_id u32 = 200017
llama_model_loader: - kv 40: tokenizer.chat_template str = {#-\n In addition to the normal input...
llama_model_loader: - type f32: 433 tensors
llama_model_loader: - type f16: 146 tensors
llama_model_loader: - type mxfp4: 108 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type = F16
print_info: file size = 60.87 GiB (4.48 BPW)
load: printing all EOG tokens:
load: - 199999 ('<|endoftext|>')
load: - 200002 ('<|return|>')
load: - 200007 ('<|end|>')
load: - 200012 ('<|call|>')
load: special_eog_ids contains both '<|return|>' and '<|call|>' tokens, removing '<|end|>' token from EOG list
load: special tokens cache size = 21
load: token to piece cache size = 1.3332 MB
print_info: arch = gpt-oss
print_info: vocab_only = 0
print_info: n_ctx_train = 131072
print_info: n_embd = 2880
print_info: n_layer = 36
print_info: n_head = 64
print_info: n_head_kv = 8
print_info: n_rot = 64
print_info: n_swa = 128
print_info: is_swa_any = 1
print_info: n_embd_head_k = 64
print_info: n_embd_head_v = 64
print_info: n_gqa = 8
print_info: n_embd_k_gqa = 512
print_info: n_embd_v_gqa = 512
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 = 2880
print_info: n_expert = 128
print_info: n_expert_used = 4
print_info: causal attn = 1
print_info: pooling type = 0
print_info: rope type = 2
print_info: rope scaling = yarn
print_info: freq_base_train = 150000.0
print_info: freq_scale_train = 0.03125
print_info: n_ctx_orig_yarn = 4096
print_info: rope_finetuned = unknown
print_info: model type = ?B
print_info: model params = 116.83 B
print_info: general.name = Gpt-Oss-120B
print_info: n_ff_exp = 2880
print_info: vocab type = BPE
print_info: n_vocab = 201088
print_info: n_merges = 446189
print_info: BOS token = 199998 '<|startoftext|>'
print_info: EOS token = 200002 '<|return|>'
print_info: EOT token = 199999 '<|endoftext|>'
print_info: PAD token = 200017 '<|reserved_200017|>'
print_info: LF token = 198 '─è'
print_info: EOG token = 199999 '<|endoftext|>'
print_info: EOG token = 200002 '<|return|>'
print_info: EOG token = 200012 '<|call|>'
print_info: max token length = 256
load_tensors: loading model tensors, this can take a while... (mmap = true)
load_tensors: offloading 11 repeating layers to GPU
load_tensors: offloaded 11/37 layers to GPU
load_tensors: Vulkan1 model buffer size = 18369.73 MiB
load_tensors: CPU_Mapped model buffer size = 54259.97 MiB
..................................................................................................
llama_context: constructing llama_context
llama_context: n_seq_max = 1
llama_context: n_ctx = 16384
llama_context: n_ctx_per_seq = 16384
llama_context: n_batch = 2048
llama_context: n_ubatch = 512
llama_context: causal_attn = 1
llama_context: flash_attn = 0
llama_context: kv_unified = false
llama_context: freq_base = 150000.0
llama_context: freq_scale = 0.03125
llama_context: n_ctx_per_seq (16384) < n_ctx_train (131072) -- the full capacity of the model will not be utilized
llama_context: CPU output buffer size = 0.77 MiB
llama_kv_cache_unified_iswa: creating non-SWA KV cache, size = 16384 cells
llama_kv_cache_unified: Vulkan1 KV buffer size = 192.00 MiB
llama_kv_cache_unified: CPU KV buffer size = 384.00 MiB
llama_kv_cache_unified: size = 576.00 MiB ( 16384 cells, 18 layers, 1/1 seqs), K (f16): 288.00 MiB, V (f16): 288.00 MiB
llama_kv_cache_unified_iswa: creating SWA KV cache, size = 640 cells
llama_kv_cache_unified: Vulkan1 KV buffer size = 6.25 MiB
llama_kv_cache_unified: CPU KV buffer size = 16.25 MiB
llama_kv_cache_unified: size = 22.50 MiB ( 640 cells, 18 layers, 1/1 seqs), K (f16): 11.25 MiB, V (f16): 11.25 MiB
ggml_vulkan: Device memory allocation of size 2221155856 failed.
ggml_vulkan: Requested buffer size exceeds device memory allocation limit: ErrorOutOfDeviceMemory
ggml_gallocr_reserve_n: failed to allocate Vulkan1 buffer of size 2221155856
graph_reserve: failed to allocate compute buffers
llama_init_from_model: failed to initialize the context: failed to allocate compute pp buffers
common_init_from_params: failed to create context with model 'gpt-oss-120b-F16.gguf'
srv load_model: failed to load model, 'gpt-oss-120b-F16.gguf'
srv operator(): operator(): cleaning up before exit...
main: exiting due to model loading error
C:\AI>vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 8 --main-gpu 1 --tensor-split 0,1
load_backend: loaded RPC backend from C:\AI\vulkan\ggml-rpc.dll
ggml_vulkan: Found 2 Vulkan devices:
ggml_vulkan: 0 = NVIDIA GeForce RTX 4060 Ti (NVIDIA) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 49152 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 1 = AMD Radeon RX 7900 XTX (AMD proprietary driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
load_backend: loaded Vulkan backend from C:\AI\vulkan\ggml-vulkan.dll
load_backend: loaded CPU backend from C:\AI\vulkan\ggml-cpu-icelake.dll
build: 6096 (fd1234cb) with clang version 19.1.5 for x86_64-pc-windows-msvc
system info: n_threads = 6, n_threads_batch = 6, total_threads = 12
system_info: n_threads = 6 (n_threads_batch = 6) / 12 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | BMI2 = 1 | AVX512 = 1 | AVX512_VBMI = 1 | AVX512_VNNI = 1 | LLAMAFILE = 1 | OPENMP = 1 | REPACK = 1 |
main: binding port with default address family
main: HTTP server is listening, hostname: 127.0.0.1, port: 8080, http threads: 11
main: loading model
srv load_model: loading model 'gpt-oss-120b-F16.gguf'
llama_model_load_from_file_impl: using device Vulkan0 (NVIDIA GeForce RTX 4060 Ti) - 16109 MiB free
llama_model_load_from_file_impl: using device Vulkan1 (AMD Radeon RX 7900 XTX) - 24560 MiB free
llama_model_loader: loaded meta data with 41 key-value pairs and 687 tensors from gpt-oss-120b-F16.gguf (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 = gpt-oss
llama_model_loader: - kv 1: general.type str = model
llama_model_loader: - kv 2: general.name str = Gpt-Oss-120B
llama_model_loader: - kv 3: general.basename str = Gpt-Oss-120B
llama_model_loader: - kv 4: general.quantized_by str = Unsloth
llama_model_loader: - kv 5: general.size_label str = 120B
llama_model_loader: - kv 6: general.license str = apache-2.0
llama_model_loader: - kv 7: general.repo_url str = https://huggingface.co/unsloth
llama_model_loader: - kv 8: general.base_model.count u32 = 1
llama_model_loader: - kv 9: general.base_model.0.name str = Gpt Oss 120b
llama_model_loader: - kv 10: general.base_model.0.organization str = Openai
llama_model_loader: - kv 11: general.base_model.0.repo_url str = https://huggingface.co/openai/gpt-oss...
llama_model_loader: - kv 12: general.tags arr[str,3] = ["openai", "unsloth", "text-generation"]
llama_model_loader: - kv 13: gpt-oss.block_count u32 = 36
llama_model_loader: - kv 14: gpt-oss.context_length u32 = 131072
llama_model_loader: - kv 15: gpt-oss.embedding_length u32 = 2880
llama_model_loader: - kv 16: gpt-oss.feed_forward_length u32 = 2880
llama_model_loader: - kv 17: gpt-oss.attention.head_count u32 = 64
llama_model_loader: - kv 18: gpt-oss.attention.head_count_kv u32 = 8
llama_model_loader: - kv 19: gpt-oss.rope.freq_base f32 = 150000.000000
llama_model_loader: - kv 20: gpt-oss.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 21: gpt-oss.expert_count u32 = 128
llama_model_loader: - kv 22: gpt-oss.expert_used_count u32 = 4
llama_model_loader: - kv 23: gpt-oss.attention.key_length u32 = 64
llama_model_loader: - kv 24: gpt-oss.attention.value_length u32 = 64
llama_model_loader: - kv 25: general.file_type u32 = 1
llama_model_loader: - kv 26: gpt-oss.attention.sliding_window u32 = 128
llama_model_loader: - kv 27: gpt-oss.expert_feed_forward_length u32 = 2880
llama_model_loader: - kv 28: gpt-oss.rope.scaling.type str = yarn
llama_model_loader: - kv 29: gpt-oss.rope.scaling.factor f32 = 32.000000
llama_model_loader: - kv 30: gpt-oss.rope.scaling.original_context_length u32 = 4096
llama_model_loader: - kv 31: general.quantization_version u32 = 2
llama_model_loader: - kv 32: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 33: tokenizer.ggml.pre str = gpt-4o
llama_model_loader: - kv 34: tokenizer.ggml.tokens arr[str,201088] = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 35: tokenizer.ggml.token_type arr[i32,201088] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 36: tokenizer.ggml.merges arr[str,446189] = ["─á ─á", "─á ─á─á─á", "─á─á ─á─á", "...
llama_model_loader: - kv 37: tokenizer.ggml.bos_token_id u32 = 199998
llama_model_loader: - kv 38: tokenizer.ggml.eos_token_id u32 = 200002
llama_model_loader: - kv 39: tokenizer.ggml.padding_token_id u32 = 200017
llama_model_loader: - kv 40: tokenizer.chat_template str = {#-\n In addition to the normal input...
llama_model_loader: - type f32: 433 tensors
llama_model_loader: - type f16: 146 tensors
llama_model_loader: - type mxfp4: 108 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type = F16
print_info: file size = 60.87 GiB (4.48 BPW)
load: printing all EOG tokens:
load: - 199999 ('<|endoftext|>')
load: - 200002 ('<|return|>')
load: - 200007 ('<|end|>')
load: - 200012 ('<|call|>')
load: special_eog_ids contains both '<|return|>' and '<|call|>' tokens, removing '<|end|>' token from EOG list
load: special tokens cache size = 21
load: token to piece cache size = 1.3332 MB
print_info: arch = gpt-oss
print_info: vocab_only = 0
print_info: n_ctx_train = 131072
print_info: n_embd = 2880
print_info: n_layer = 36
print_info: n_head = 64
print_info: n_head_kv = 8
print_info: n_rot = 64
print_info: n_swa = 128
print_info: is_swa_any = 1
print_info: n_embd_head_k = 64
print_info: n_embd_head_v = 64
print_info: n_gqa = 8
print_info: n_embd_k_gqa = 512
print_info: n_embd_v_gqa = 512
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 = 2880
print_info: n_expert = 128
print_info: n_expert_used = 4
print_info: causal attn = 1
print_info: pooling type = 0
print_info: rope type = 2
print_info: rope scaling = yarn
print_info: freq_base_train = 150000.0
print_info: freq_scale_train = 0.03125
print_info: n_ctx_orig_yarn = 4096
print_info: rope_finetuned = unknown
print_info: model type = ?B
print_info: model params = 116.83 B
print_info: general.name = Gpt-Oss-120B
print_info: n_ff_exp = 2880
print_info: vocab type = BPE
print_info: n_vocab = 201088
print_info: n_merges = 446189
print_info: BOS token = 199998 '<|startoftext|>'
print_info: EOS token = 200002 '<|return|>'
print_info: EOT token = 199999 '<|endoftext|>'
print_info: PAD token = 200017 '<|reserved_200017|>'
print_info: LF token = 198 '─è'
print_info: EOG token = 199999 '<|endoftext|>'
print_info: EOG token = 200002 '<|return|>'
print_info: EOG token = 200012 '<|call|>'
print_info: max token length = 256
load_tensors: loading model tensors, this can take a while... (mmap = true)
load_tensors: offloading 8 repeating layers to GPU
load_tensors: offloaded 8/37 layers to GPU
load_tensors: Vulkan1 model buffer size = 13359.80 MiB
load_tensors: CPU_Mapped model buffer size = 62328.33 MiB
...................................................................................................
llama_context: constructing llama_context
llama_context: n_seq_max = 1
llama_context: n_ctx = 16384
llama_context: n_ctx_per_seq = 16384
llama_context: n_batch = 2048
llama_context: n_ubatch = 512
llama_context: causal_attn = 1
llama_context: flash_attn = 0
llama_context: kv_unified = false
llama_context: freq_base = 150000.0
llama_context: freq_scale = 0.03125
llama_context: n_ctx_per_seq (16384) < n_ctx_train (131072) -- the full capacity of the model will not be utilized
llama_context: CPU output buffer size = 0.77 MiB
llama_kv_cache_unified_iswa: creating non-SWA KV cache, size = 16384 cells
llama_kv_cache_unified: Vulkan1 KV buffer size = 128.00 MiB
llama_kv_cache_unified: CPU KV buffer size = 448.00 MiB
llama_kv_cache_unified: size = 576.00 MiB ( 16384 cells, 18 layers, 1/1 seqs), K (f16): 288.00 MiB, V (f16): 288.00 MiB
llama_kv_cache_unified_iswa: creating SWA KV cache, size = 640 cells
llama_kv_cache_unified: Vulkan1 KV buffer size = 5.00 MiB
llama_kv_cache_unified: CPU KV buffer size = 17.50 MiB
llama_kv_cache_unified: size = 22.50 MiB ( 640 cells, 18 layers, 1/1 seqs), K (f16): 11.25 MiB, V (f16): 11.25 MiB
ggml_vulkan: Device memory allocation of size 2216568336 failed.
ggml_vulkan: Requested buffer size exceeds device memory allocation limit: ErrorOutOfDeviceMemory
ggml_gallocr_reserve_n: failed to allocate Vulkan1 buffer of size 2216568336
graph_reserve: failed to allocate compute buffers
llama_init_from_model: failed to initialize the context: failed to allocate compute pp buffers
common_init_from_params: failed to create context with model 'gpt-oss-120b-F16.gguf'
srv load_model: failed to load model, 'gpt-oss-120b-F16.gguf'
srv operator(): operator(): cleaning up before exit...
main: exiting due to model loading error
C:\AI>vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 16 --main-gpu 0 --tensor-split 6,10
load_backend: loaded RPC backend from C:\AI\vulkan\ggml-rpc.dll
ggml_vulkan: Found 2 Vulkan devices:
ggml_vulkan: 0 = NVIDIA GeForce RTX 4060 Ti (NVIDIA) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 49152 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 1 = AMD Radeon RX 7900 XTX (AMD proprietary driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
load_backend: loaded Vulkan backend from C:\AI\vulkan\ggml-vulkan.dll
load_backend: loaded CPU backend from C:\AI\vulkan\ggml-cpu-icelake.dll
build: 6096 (fd1234cb) with clang version 19.1.5 for x86_64-pc-windows-msvc
system info: n_threads = 6, n_threads_batch = 6, total_threads = 12
system_info: n_threads = 6 (n_threads_batch = 6) / 12 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | BMI2 = 1 | AVX512 = 1 | AVX512_VBMI = 1 | AVX512_VNNI = 1 | LLAMAFILE = 1 | OPENMP = 1 | REPACK = 1 |
main: binding port with default address family
main: HTTP server is listening, hostname: 127.0.0.1, port: 8080, http threads: 11
main: loading model
srv load_model: loading model 'gpt-oss-120b-F16.gguf'
llama_model_load_from_file_impl: using device Vulkan0 (NVIDIA GeForce RTX 4060 Ti) - 16109 MiB free
llama_model_load_from_file_impl: using device Vulkan1 (AMD Radeon RX 7900 XTX) - 24560 MiB free
llama_model_loader: loaded meta data with 41 key-value pairs and 687 tensors from gpt-oss-120b-F16.gguf (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 = gpt-oss
llama_model_loader: - kv 1: general.type str = model
llama_model_loader: - kv 2: general.name str = Gpt-Oss-120B
llama_model_loader: - kv 3: general.basename str = Gpt-Oss-120B
llama_model_loader: - kv 4: general.quantized_by str = Unsloth
llama_model_loader: - kv 5: general.size_label str = 120B
llama_model_loader: - kv 6: general.license str = apache-2.0
llama_model_loader: - kv 7: general.repo_url str = https://huggingface.co/unsloth
llama_model_loader: - kv 8: general.base_model.count u32 = 1
llama_model_loader: - kv 9: general.base_model.0.name str = Gpt Oss 120b
llama_model_loader: - kv 10: general.base_model.0.organization str = Openai
llama_model_loader: - kv 11: general.base_model.0.repo_url str = https://huggingface.co/openai/gpt-oss...
llama_model_loader: - kv 12: general.tags arr[str,3] = ["openai", "unsloth", "text-generation"]
llama_model_loader: - kv 13: gpt-oss.block_count u32 = 36
llama_model_loader: - kv 14: gpt-oss.context_length u32 = 131072
llama_model_loader: - kv 15: gpt-oss.embedding_length u32 = 2880
llama_model_loader: - kv 16: gpt-oss.feed_forward_length u32 = 2880
llama_model_loader: - kv 17: gpt-oss.attention.head_count u32 = 64
llama_model_loader: - kv 18: gpt-oss.attention.head_count_kv u32 = 8
llama_model_loader: - kv 19: gpt-oss.rope.freq_base f32 = 150000.000000
llama_model_loader: - kv 20: gpt-oss.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 21: gpt-oss.expert_count u32 = 128
llama_model_loader: - kv 22: gpt-oss.expert_used_count u32 = 4
llama_model_loader: - kv 23: gpt-oss.attention.key_length u32 = 64
llama_model_loader: - kv 24: gpt-oss.attention.value_length u32 = 64
llama_model_loader: - kv 25: general.file_type u32 = 1
llama_model_loader: - kv 26: gpt-oss.attention.sliding_window u32 = 128
llama_model_loader: - kv 27: gpt-oss.expert_feed_forward_length u32 = 2880
llama_model_loader: - kv 28: gpt-oss.rope.scaling.type str = yarn
llama_model_loader: - kv 29: gpt-oss.rope.scaling.factor f32 = 32.000000
llama_model_loader: - kv 30: gpt-oss.rope.scaling.original_context_length u32 = 4096
llama_model_loader: - kv 31: general.quantization_version u32 = 2
llama_model_loader: - kv 32: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 33: tokenizer.ggml.pre str = gpt-4o
llama_model_loader: - kv 34: tokenizer.ggml.tokens arr[str,201088] = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 35: tokenizer.ggml.token_type arr[i32,201088] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 36: tokenizer.ggml.merges arr[str,446189] = ["─á ─á", "─á ─á─á─á", "─á─á ─á─á", "...
llama_model_loader: - kv 37: tokenizer.ggml.bos_token_id u32 = 199998
llama_model_loader: - kv 38: tokenizer.ggml.eos_token_id u32 = 200002
llama_model_loader: - kv 39: tokenizer.ggml.padding_token_id u32 = 200017
llama_model_loader: - kv 40: tokenizer.chat_template str = {#-\n In addition to the normal input...
llama_model_loader: - type f32: 433 tensors
llama_model_loader: - type f16: 146 tensors
llama_model_loader: - type mxfp4: 108 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type = F16
print_info: file size = 60.87 GiB (4.48 BPW)
load: printing all EOG tokens:
load: - 199999 ('<|endoftext|>')
load: - 200002 ('<|return|>')
load: - 200007 ('<|end|>')
load: - 200012 ('<|call|>')
load: special_eog_ids contains both '<|return|>' and '<|call|>' tokens, removing '<|end|>' token from EOG list
load: special tokens cache size = 21
load: token to piece cache size = 1.3332 MB
print_info: arch = gpt-oss
print_info: vocab_only = 0
print_info: n_ctx_train = 131072
print_info: n_embd = 2880
print_info: n_layer = 36
print_info: n_head = 64
print_info: n_head_kv = 8
print_info: n_rot = 64
print_info: n_swa = 128
print_info: is_swa_any = 1
print_info: n_embd_head_k = 64
print_info: n_embd_head_v = 64
print_info: n_gqa = 8
print_info: n_embd_k_gqa = 512
print_info: n_embd_v_gqa = 512
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 = 2880
print_info: n_expert = 128
print_info: n_expert_used = 4
print_info: causal attn = 1
print_info: pooling type = 0
print_info: rope type = 2
print_info: rope scaling = yarn
print_info: freq_base_train = 150000.0
print_info: freq_scale_train = 0.03125
print_info: n_ctx_orig_yarn = 4096
print_info: rope_finetuned = unknown
print_info: model type = ?B
print_info: model params = 116.83 B
print_info: general.name = Gpt-Oss-120B
print_info: n_ff_exp = 2880
print_info: vocab type = BPE
print_info: n_vocab = 201088
print_info: n_merges = 446189
print_info: BOS token = 199998 '<|startoftext|>'
print_info: EOS token = 200002 '<|return|>'
print_info: EOT token = 199999 '<|endoftext|>'
print_info: PAD token = 200017 '<|reserved_200017|>'
print_info: LF token = 198 '─è'
print_info: EOG token = 199999 '<|endoftext|>'
print_info: EOG token = 200002 '<|return|>'
print_info: EOG token = 200012 '<|call|>'
print_info: max token length = 256
load_tensors: loading model tensors, this can take a while... (mmap = true)
load_tensors: offloading 16 repeating layers to GPU
load_tensors: offloaded 16/37 layers to GPU
load_tensors: Vulkan0 model buffer size = 10019.85 MiB
load_tensors: Vulkan1 model buffer size = 16699.75 MiB
load_tensors: CPU_Mapped model buffer size = 54259.97 MiB
..................................................................................................
llama_context: constructing llama_context
llama_context: n_seq_max = 1
llama_context: n_ctx = 16384
llama_context: n_ctx_per_seq = 16384
llama_context: n_batch = 2048
llama_context: n_ubatch = 512
llama_context: causal_attn = 1
llama_context: flash_attn = 0
llama_context: kv_unified = false
llama_context: freq_base = 150000.0
llama_context: freq_scale = 0.03125
llama_context: n_ctx_per_seq (16384) < n_ctx_train (131072) -- the full capacity of the model will not be utilized
llama_context: CPU output buffer size = 0.77 MiB
llama_kv_cache_unified_iswa: creating non-SWA KV cache, size = 16384 cells
llama_kv_cache_unified: Vulkan0 KV buffer size = 96.00 MiB
llama_kv_cache_unified: Vulkan1 KV buffer size = 160.00 MiB
llama_kv_cache_unified: CPU KV buffer size = 320.00 MiB
llama_kv_cache_unified: size = 576.00 MiB ( 16384 cells, 18 layers, 1/1 seqs), K (f16): 288.00 MiB, V (f16): 288.00 MiB
llama_kv_cache_unified_iswa: creating SWA KV cache, size = 640 cells
llama_kv_cache_unified: Vulkan0 KV buffer size = 3.75 MiB
llama_kv_cache_unified: Vulkan1 KV buffer size = 6.25 MiB
llama_kv_cache_unified: CPU KV buffer size = 12.50 MiB
llama_kv_cache_unified: size = 22.50 MiB ( 640 cells, 18 layers, 1/1 seqs), K (f16): 11.25 MiB, V (f16): 11.25 MiB
ggml_vulkan: Device memory allocation of size 2215129088 failed.
ggml_vulkan: Requested buffer size exceeds device memory allocation limit: ErrorOutOfDeviceMemory
ggml_gallocr_reserve_n: failed to allocate Vulkan1 buffer of size 2215129088
graph_reserve: failed to allocate compute buffers
llama_init_from_model: failed to initialize the context: failed to allocate compute pp buffers
common_init_from_params: failed to create context with model 'gpt-oss-120b-F16.gguf'
srv load_model: failed to load model, 'gpt-oss-120b-F16.gguf'
srv operator(): operator(): cleaning up before exit...
main: exiting due to model loading error
C:\AI>vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 7 --main-gpu 0 --tensor-split 1,0
load_backend: loaded RPC backend from C:\AI\vulkan\ggml-rpc.dll
ggml_vulkan: Found 2 Vulkan devices:
ggml_vulkan: 0 = NVIDIA GeForce RTX 4060 Ti (NVIDIA) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 49152 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 1 = AMD Radeon RX 7900 XTX (AMD proprietary driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
load_backend: loaded Vulkan backend from C:\AI\vulkan\ggml-vulkan.dll
load_backend: loaded CPU backend from C:\AI\vulkan\ggml-cpu-icelake.dll
build: 6096 (fd1234cb) with clang version 19.1.5 for x86_64-pc-windows-msvc
system info: n_threads = 6, n_threads_batch = 6, total_threads = 12
system_info: n_threads = 6 (n_threads_batch = 6) / 12 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | BMI2 = 1 | AVX512 = 1 | AVX512_VBMI = 1 | AVX512_VNNI = 1 | LLAMAFILE = 1 | OPENMP = 1 | REPACK = 1 |
main: binding port with default address family
main: HTTP server is listening, hostname: 127.0.0.1, port: 8080, http threads: 11
main: loading model
srv load_model: loading model 'gpt-oss-120b-F16.gguf'
llama_model_load_from_file_impl: using device Vulkan0 (NVIDIA GeForce RTX 4060 Ti) - 16109 MiB free
llama_model_load_from_file_impl: using device Vulkan1 (AMD Radeon RX 7900 XTX) - 24560 MiB free
llama_model_loader: loaded meta data with 41 key-value pairs and 687 tensors from gpt-oss-120b-F16.gguf (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 = gpt-oss
llama_model_loader: - kv 1: general.type str = model
llama_model_loader: - kv 2: general.name str = Gpt-Oss-120B
llama_model_loader: - kv 3: general.basename str = Gpt-Oss-120B
llama_model_loader: - kv 4: general.quantized_by str = Unsloth
llama_model_loader: - kv 5: general.size_label str = 120B
llama_model_loader: - kv 6: general.license str = apache-2.0
llama_model_loader: - kv 7: general.repo_url str = https://huggingface.co/unsloth
llama_model_loader: - kv 8: general.base_model.count u32 = 1
llama_model_loader: - kv 9: general.base_model.0.name str = Gpt Oss 120b
llama_model_loader: - kv 10: general.base_model.0.organization str = Openai
llama_model_loader: - kv 11: general.base_model.0.repo_url str = https://huggingface.co/openai/gpt-oss...
llama_model_loader: - kv 12: general.tags arr[str,3] = ["openai", "unsloth", "text-generation"]
llama_model_loader: - kv 13: gpt-oss.block_count u32 = 36
llama_model_loader: - kv 14: gpt-oss.context_length u32 = 131072
llama_model_loader: - kv 15: gpt-oss.embedding_length u32 = 2880
llama_model_loader: - kv 16: gpt-oss.feed_forward_length u32 = 2880
llama_model_loader: - kv 17: gpt-oss.attention.head_count u32 = 64
llama_model_loader: - kv 18: gpt-oss.attention.head_count_kv u32 = 8
llama_model_loader: - kv 19: gpt-oss.rope.freq_base f32 = 150000.000000
llama_model_loader: - kv 20: gpt-oss.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 21: gpt-oss.expert_count u32 = 128
llama_model_loader: - kv 22: gpt-oss.expert_used_count u32 = 4
llama_model_loader: - kv 23: gpt-oss.attention.key_length u32 = 64
llama_model_loader: - kv 24: gpt-oss.attention.value_length u32 = 64
llama_model_loader: - kv 25: general.file_type u32 = 1
llama_model_loader: - kv 26: gpt-oss.attention.sliding_window u32 = 128
llama_model_loader: - kv 27: gpt-oss.expert_feed_forward_length u32 = 2880
llama_model_loader: - kv 28: gpt-oss.rope.scaling.type str = yarn
llama_model_loader: - kv 29: gpt-oss.rope.scaling.factor f32 = 32.000000
llama_model_loader: - kv 30: gpt-oss.rope.scaling.original_context_length u32 = 4096
llama_model_loader: - kv 31: general.quantization_version u32 = 2
llama_model_loader: - kv 32: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 33: tokenizer.ggml.pre str = gpt-4o
llama_model_loader: - kv 34: tokenizer.ggml.tokens arr[str,201088] = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 35: tokenizer.ggml.token_type arr[i32,201088] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 36: tokenizer.ggml.merges arr[str,446189] = ["─á ─á", "─á ─á─á─á", "─á─á ─á─á", "...
llama_model_loader: - kv 37: tokenizer.ggml.bos_token_id u32 = 199998
llama_model_loader: - kv 38: tokenizer.ggml.eos_token_id u32 = 200002
llama_model_loader: - kv 39: tokenizer.ggml.padding_token_id u32 = 200017
llama_model_loader: - kv 40: tokenizer.chat_template str = {#-\n In addition to the normal input...
llama_model_loader: - type f32: 433 tensors
llama_model_loader: - type f16: 146 tensors
llama_model_loader: - type mxfp4: 108 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type = F16
print_info: file size = 60.87 GiB (4.48 BPW)
load: printing all EOG tokens:
load: - 199999 ('<|endoftext|>')
load: - 200002 ('<|return|>')
load: - 200007 ('<|end|>')
load: - 200012 ('<|call|>')
load: special_eog_ids contains both '<|return|>' and '<|call|>' tokens, removing '<|end|>' token from EOG list
load: special tokens cache size = 21
load: token to piece cache size = 1.3332 MB
print_info: arch = gpt-oss
print_info: vocab_only = 0
print_info: n_ctx_train = 131072
print_info: n_embd = 2880
print_info: n_layer = 36
print_info: n_head = 64
print_info: n_head_kv = 8
print_info: n_rot = 64
print_info: n_swa = 128
print_info: is_swa_any = 1
print_info: n_embd_head_k = 64
print_info: n_embd_head_v = 64
print_info: n_gqa = 8
print_info: n_embd_k_gqa = 512
print_info: n_embd_v_gqa = 512
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 = 2880
print_info: n_expert = 128
print_info: n_expert_used = 4
print_info: causal attn = 1
print_info: pooling type = 0
print_info: rope type = 2
print_info: rope scaling = yarn
print_info: freq_base_train = 150000.0
print_info: freq_scale_train = 0.03125
print_info: n_ctx_orig_yarn = 4096
print_info: rope_finetuned = unknown
print_info: model type = ?B
print_info: model params = 116.83 B
print_info: general.name = Gpt-Oss-120B
print_info: n_ff_exp = 2880
print_info: vocab type = BPE
print_info: n_vocab = 201088
print_info: n_merges = 446189
print_info: BOS token = 199998 '<|startoftext|>'
print_info: EOS token = 200002 '<|return|>'
print_info: EOT token = 199999 '<|endoftext|>'
print_info: PAD token = 200017 '<|reserved_200017|>'
print_info: LF token = 198 '─è'
print_info: EOG token = 199999 '<|endoftext|>'
print_info: EOG token = 200002 '<|return|>'
print_info: EOG token = 200012 '<|call|>'
print_info: max token length = 256
load_tensors: loading model tensors, this can take a while... (mmap = true)
load_tensors: offloading 7 repeating layers to GPU
load_tensors: offloaded 7/37 layers to GPU
load_tensors: Vulkan0 model buffer size = 11689.83 MiB
load_tensors: CPU_Mapped model buffer size = 62328.33 MiB
..................................................................................................
llama_context: constructing llama_context
llama_context: n_seq_max = 1
llama_context: n_ctx = 16384
llama_context: n_ctx_per_seq = 16384
llama_context: n_batch = 2048
llama_context: n_ubatch = 512
llama_context: causal_attn = 1
llama_context: flash_attn = 0
llama_context: kv_unified = false
llama_context: freq_base = 150000.0
llama_context: freq_scale = 0.03125
llama_context: n_ctx_per_seq (16384) < n_ctx_train (131072) -- the full capacity of the model will not be utilized
llama_context: CPU output buffer size = 0.77 MiB
llama_kv_cache_unified_iswa: creating non-SWA KV cache, size = 16384 cells
llama_kv_cache_unified: Vulkan0 KV buffer size = 128.00 MiB
llama_kv_cache_unified: CPU KV buffer size = 448.00 MiB
llama_kv_cache_unified: size = 576.00 MiB ( 16384 cells, 18 layers, 1/1 seqs), K (f16): 288.00 MiB, V (f16): 288.00 MiB
llama_kv_cache_unified_iswa: creating SWA KV cache, size = 640 cells
llama_kv_cache_unified: Vulkan0 KV buffer size = 3.75 MiB
llama_kv_cache_unified: CPU KV buffer size = 18.75 MiB
llama_kv_cache_unified: size = 22.50 MiB ( 640 cells, 18 layers, 1/1 seqs), K (f16): 11.25 MiB, V (f16): 11.25 MiB
ggml_vulkan: Device memory allocation of size 4371400704 failed.
ggml_vulkan: Requested buffer size exceeds device memory allocation limit: ErrorOutOfDeviceMemory
ggml_gallocr_reserve_n: failed to allocate Vulkan0 buffer of size 4371400704
graph_reserve: failed to allocate compute buffers
llama_init_from_model: failed to initialize the context: failed to allocate compute pp buffers
common_init_from_params: failed to create context with model 'gpt-oss-120b-F16.gguf'
srv load_model: failed to load model, 'gpt-oss-120b-F16.gguf'
srv operator(): operator(): cleaning up before exit...
main: exiting due to model loading error
C:\AI>vulkan\llama-server -m gpt-oss-120b-F16.gguf -c 16384 -ngl 6 --main-gpu 0 --tensor-split 1,0
load_backend: loaded RPC backend from C:\AI\vulkan\ggml-rpc.dll
ggml_vulkan: Found 2 Vulkan devices:
ggml_vulkan: 0 = NVIDIA GeForce RTX 4060 Ti (NVIDIA) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 49152 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 1 = AMD Radeon RX 7900 XTX (AMD proprietary driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
load_backend: loaded Vulkan backend from C:\AI\vulkan\ggml-vulkan.dll
load_backend: loaded CPU backend from C:\AI\vulkan\ggml-cpu-icelake.dll
build: 6096 (fd1234cb) with clang version 19.1.5 for x86_64-pc-windows-msvc
system info: n_threads = 6, n_threads_batch = 6, total_threads = 12
system_info: n_threads = 6 (n_threads_batch = 6) / 12 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | BMI2 = 1 | AVX512 = 1 | AVX512_VBMI = 1 | AVX512_VNNI = 1 | LLAMAFILE = 1 | OPENMP = 1 | REPACK = 1 |
main: binding port with default address family
main: HTTP server is listening, hostname: 127.0.0.1, port: 8080, http threads: 11
main: loading model
srv load_model: loading model 'gpt-oss-120b-F16.gguf'
llama_model_load_from_file_impl: using device Vulkan0 (NVIDIA GeForce RTX 4060 Ti) - 16109 MiB free
llama_model_load_from_file_impl: using device Vulkan1 (AMD Radeon RX 7900 XTX) - 24560 MiB free
llama_model_loader: loaded meta data with 41 key-value pairs and 687 tensors from gpt-oss-120b-F16.gguf (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 = gpt-oss
llama_model_loader: - kv 1: general.type str = model
llama_model_loader: - kv 2: general.name str = Gpt-Oss-120B
llama_model_loader: - kv 3: general.basename str = Gpt-Oss-120B
llama_model_loader: - kv 4: general.quantized_by str = Unsloth
llama_model_loader: - kv 5: general.size_label str = 120B
llama_model_loader: - kv 6: general.license str = apache-2.0
llama_model_loader: - kv 7: general.repo_url str = https://huggingface.co/unsloth
llama_model_loader: - kv 8: general.base_model.count u32 = 1
llama_model_loader: - kv 9: general.base_model.0.name str = Gpt Oss 120b
llama_model_loader: - kv 10: general.base_model.0.organization str = Openai
llama_model_loader: - kv 11: general.base_model.0.repo_url str = https://huggingface.co/openai/gpt-oss...
llama_model_loader: - kv 12: general.tags arr[str,3] = ["openai", "unsloth", "text-generation"]
llama_model_loader: - kv 13: gpt-oss.block_count u32 = 36
llama_model_loader: - kv 14: gpt-oss.context_length u32 = 131072
llama_model_loader: - kv 15: gpt-oss.embedding_length u32 = 2880
llama_model_loader: - kv 16: gpt-oss.feed_forward_length u32 = 2880
llama_model_loader: - kv 17: gpt-oss.attention.head_count u32 = 64
llama_model_loader: - kv 18: gpt-oss.attention.head_count_kv u32 = 8
llama_model_loader: - kv 19: gpt-oss.rope.freq_base f32 = 150000.000000
llama_model_loader: - kv 20: gpt-oss.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 21: gpt-oss.expert_count u32 = 128
llama_model_loader: - kv 22: gpt-oss.expert_used_count u32 = 4
llama_model_loader: - kv 23: gpt-oss.attention.key_length u32 = 64
llama_model_loader: - kv 24: gpt-oss.attention.value_length u32 = 64
llama_model_loader: - kv 25: general.file_type u32 = 1
llama_model_loader: - kv 26: gpt-oss.attention.sliding_window u32 = 128
llama_model_loader: - kv 27: gpt-oss.expert_feed_forward_length u32 = 2880
llama_model_loader: - kv 28: gpt-oss.rope.scaling.type str = yarn
llama_model_loader: - kv 29: gpt-oss.rope.scaling.factor f32 = 32.000000
llama_model_loader: - kv 30: gpt-oss.rope.scaling.original_context_length u32 = 4096
llama_model_loader: - kv 31: general.quantization_version u32 = 2
llama_model_loader: - kv 32: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 33: tokenizer.ggml.pre str = gpt-4o
llama_model_loader: - kv 34: tokenizer.ggml.tokens arr[str,201088] = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 35: tokenizer.ggml.token_type arr[i32,201088] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 36: tokenizer.ggml.merges arr[str,446189] = ["─á ─á", "─á ─á─á─á", "─á─á ─á─á", "...
llama_model_loader: - kv 37: tokenizer.ggml.bos_token_id u32 = 199998
llama_model_loader: - kv 38: tokenizer.ggml.eos_token_id u32 = 200002
llama_model_loader: - kv 39: tokenizer.ggml.padding_token_id u32 = 200017
llama_model_loader: - kv 40: tokenizer.chat_template str = {#-\n In addition to the normal input...
llama_model_loader: - type f32: 433 tensors
llama_model_loader: - type f16: 146 tensors
llama_model_loader: - type mxfp4: 108 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type = F16
print_info: file size = 60.87 GiB (4.48 BPW)
load: printing all EOG tokens:
load: - 199999 ('<|endoftext|>')
load: - 200002 ('<|return|>')
load: - 200007 ('<|end|>')
load: - 200012 ('<|call|>')
load: special_eog_ids contains both '<|return|>' and '<|call|>' tokens, removing '<|end|>' token from EOG list
load: special tokens cache size = 21
load: token to piece cache size = 1.3332 MB
print_info: arch = gpt-oss
print_info: vocab_only = 0
print_info: n_ctx_train = 131072
print_info: n_embd = 2880
print_info: n_layer = 36
print_info: n_head = 64
print_info: n_head_kv = 8
print_info: n_rot = 64
print_info: n_swa = 128
print_info: is_swa_any = 1
print_info: n_embd_head_k = 64
print_info: n_embd_head_v = 64
print_info: n_gqa = 8
print_info: n_embd_k_gqa = 512
print_info: n_embd_v_gqa = 512
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 = 2880
print_info: n_expert = 128
print_info: n_expert_used = 4
print_info: causal attn = 1
print_info: pooling type = 0
print_info: rope type = 2
print_info: rope scaling = yarn
print_info: freq_base_train = 150000.0
print_info: freq_scale_train = 0.03125
print_info: n_ctx_orig_yarn = 4096
print_info: rope_finetuned = unknown
print_info: model type = ?B
print_info: model params = 116.83 B
print_info: general.name = Gpt-Oss-120B
print_info: n_ff_exp = 2880
print_info: vocab type = BPE
print_info: n_vocab = 201088
print_info: n_merges = 446189
print_info: BOS token = 199998 '<|startoftext|>'
print_info: EOS token = 200002 '<|return|>'
print_info: EOT token = 199999 '<|endoftext|>'
print_info: PAD token = 200017 '<|reserved_200017|>'
print_info: LF token = 198 '─è'
print_info: EOG token = 199999 '<|endoftext|>'
print_info: EOG token = 200002 '<|return|>'
print_info: EOG token = 200012 '<|call|>'
print_info: max token length = 256
load_tensors: loading model tensors, this can take a while... (mmap = true)
load_tensors: offloading 6 repeating layers to GPU
load_tensors: offloaded 6/37 layers to GPU
load_tensors: Vulkan0 model buffer size = 10019.85 MiB
load_tensors: CPU_Mapped model buffer size = 62328.33 MiB
...................................................................................................
llama_context: constructing llama_context
llama_context: n_seq_max = 1
llama_context: n_ctx = 16384
llama_context: n_ctx_per_seq = 16384
llama_context: n_batch = 2048
llama_context: n_ubatch = 512
llama_context: causal_attn = 1
llama_context: flash_attn = 0
llama_context: kv_unified = false
llama_context: freq_base = 150000.0
llama_context: freq_scale = 0.03125
llama_context: n_ctx_per_seq (16384) < n_ctx_train (131072) -- the full capacity of the model will not be utilized
llama_context: CPU output buffer size = 0.77 MiB
llama_kv_cache_unified_iswa: creating non-SWA KV cache, size = 16384 cells
llama_kv_cache_unified: Vulkan0 KV buffer size = 96.00 MiB
llama_kv_cache_unified: CPU KV buffer size = 480.00 MiB
llama_kv_cache_unified: size = 576.00 MiB ( 16384 cells, 18 layers, 1/1 seqs), K (f16): 288.00 MiB, V (f16): 288.00 MiB
llama_kv_cache_unified_iswa: creating SWA KV cache, size = 640 cells
llama_kv_cache_unified: Vulkan0 KV buffer size = 3.75 MiB
llama_kv_cache_unified: CPU KV buffer size = 18.75 MiB
llama_kv_cache_unified: size = 22.50 MiB ( 640 cells, 18 layers, 1/1 seqs), K (f16): 11.25 MiB, V (f16): 11.25 MiB
llama_context: Vulkan0 compute buffer size = 2120.89 MiB
llama_context: Vulkan_Host compute buffer size = 82.26 MiB
llama_context: graph nodes = 2166
llama_context: graph splits = 694 (with bs=512), 4 (with bs=1)
common_init_from_params: KV cache shifting is not supported for this context, disabling KV cache shifting
common_init_from_params: added <|endoftext|> logit bias = -inf
common_init_from_params: added <|return|> logit bias = -inf
common_init_from_params: added <|call|> logit bias = -inf
common_init_from_params: setting dry_penalty_last_n to ctx_size = 16384
common_init_from_params: warming up the model with an empty run - please wait