Bug: error loading model DeepSeek-R1-Q4_K_M.gguf : error loading model vocabulary: unknown pre-tokenizer type: 'deepseek-v3'

[Tridu33]

Contact Details

[email protected]

What happened?

i download GGUF files DeepSeek-R1-Q4_K_M from https://www.modelscope.cn/models/unsloth/DeepSeek-R1-GGUF/files , merge them into one single file:

/home/tridu33/workspace/llama.cpp4Ascend/build/bin/llama-gguf-split --merge /home/tridu33/workspace/soft/llamafile/models/DeepSeek-R1-Q4_K_M-split/DeepSeek-R1-Q4_K_M-00001-of-00009.gguf  /home/tridu33/workspace/soft/llamafile/models/DeepSeek-R1-Q4_K_M_one/DeepSeek-R1-Q4_K_M.gguf

Version

branch main llamafile latest

Relevant log output

compile llamafile from source, then i run 👍

/home/tridu33/workspace/soft/llamafile/o/llama.cpp/main/main  -ngl 9999 -m /home/tridu33/workspace/soft/llamafile/models/DeepSeek-R1-Q4_K_M_one/DeepSeek-R1-Q4_K_M.gguf -p '[INST]Write a story about llamas[/INST]'
import_cuda_impl: initializing gpu module...
get_rocm_bin_path: note: hipcc not found on $PATH
get_rocm_bin_path: note: $HIP_PATH/bin/hipcc does not exist
get_rocm_bin_path: note: /opt/rocm/bin/hipcc does not exist
extract_cuda_dso: note: prebuilt binary /zip/ggml-rocm.so not found
import_cuda_impl: won't compile AMD GPU support because $HIP_PATH/bin/clang++ is missing
extract_cuda_dso: note: prebuilt binary /zip/ggml-rocm.so not found
get_nvcc_path: note: nvcc not found on $PATH
get_nvcc_path: note: $CUDA_PATH/bin/nvcc does not exist
get_nvcc_path: note: /opt/cuda/bin/nvcc does not exist
get_nvcc_path: note: /usr/local/cuda/bin/nvcc does not exist
extract_cuda_dso: note: prebuilt binary /zip/ggml-cuda.so not found
extract_cuda_dso: note: prebuilt binary /zip/ggml-cuda.so not found
warning: --n-gpu-layers 9999 was passed but no GPUs were found; falling back to CPU inference
Log start
main: build = 1500 (a30b324)
main: built with cosmocc (GCC) 11.2.0 for x86_64-linux-cosmo
main: seed  = 1740455235
llama_model_loader: loaded meta data with 48 key-value pairs and 1025 tensors from /home/tridu33/workspace/soft/llamafile/models/DeepSeek-R1-Q4_K_M_one/DeepSeek-R1-Q4_K_M.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              = deepseek2
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = DeepSeek R1 BF16
llama_model_loader: - kv   3:                       general.quantized_by str              = Unsloth
llama_model_loader: - kv   4:                         general.size_label str              = 256x20B
llama_model_loader: - kv   5:                           general.repo_url str              = https://huggingface.co/unsloth
llama_model_loader: - kv   6:                      deepseek2.block_count u32              = 61
llama_model_loader: - kv   7:                   deepseek2.context_length u32              = 163840
llama_model_loader: - kv   8:                 deepseek2.embedding_length u32              = 7168
llama_model_loader: - kv   9:              deepseek2.feed_forward_length u32              = 18432
llama_model_loader: - kv  10:             deepseek2.attention.head_count u32              = 128
llama_model_loader: - kv  11:          deepseek2.attention.head_count_kv u32              = 128
llama_model_loader: - kv  12:                   deepseek2.rope.freq_base f32              = 10000.000000
llama_model_loader: - kv  13: deepseek2.attention.layer_norm_rms_epsilon f32              = 0.000001
llama_model_loader: - kv  14:                deepseek2.expert_used_count u32              = 8
llama_model_loader: - kv  15:        deepseek2.leading_dense_block_count u32              = 3
llama_model_loader: - kv  16:                       deepseek2.vocab_size u32              = 129280
llama_model_loader: - kv  17:            deepseek2.attention.q_lora_rank u32              = 1536
llama_model_loader: - kv  18:           deepseek2.attention.kv_lora_rank u32              = 512
llama_model_loader: - kv  19:             deepseek2.attention.key_length u32              = 192
llama_model_loader: - kv  20:           deepseek2.attention.value_length u32              = 128
llama_model_loader: - kv  21:       deepseek2.expert_feed_forward_length u32              = 2048
llama_model_loader: - kv  22:                     deepseek2.expert_count u32              = 256
llama_model_loader: - kv  23:              deepseek2.expert_shared_count u32              = 1
llama_model_loader: - kv  24:             deepseek2.expert_weights_scale f32              = 2.500000
llama_model_loader: - kv  25:              deepseek2.expert_weights_norm bool             = true
llama_model_loader: - kv  26:               deepseek2.expert_gating_func u32              = 2
llama_model_loader: - kv  27:             deepseek2.rope.dimension_count u32              = 64
llama_model_loader: - kv  28:                deepseek2.rope.scaling.type str              = yarn
llama_model_loader: - kv  29:              deepseek2.rope.scaling.factor f32              = 40.000000
llama_model_loader: - kv  30: deepseek2.rope.scaling.original_context_length u32              = 4096
llama_model_loader: - kv  31: deepseek2.rope.scaling.yarn_log_multiplier f32              = 0.100000
llama_model_loader: - kv  32:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  33:                         tokenizer.ggml.pre str              = deepseek-v3
llama_model_loader: - kv  34:                      tokenizer.ggml.tokens arr[str,129280]  = ["<｜begin▁of▁sentence｜>", "<�...
llama_model_loader: - kv  35:                  tokenizer.ggml.token_type arr[i32,129280]  = [3, 3, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv  36:                      tokenizer.ggml.merges arr[str,127741]  = ["Ġ t", "Ġ a", "i n", "Ġ Ġ", "h e...
llama_model_loader: - kv  37:                tokenizer.ggml.bos_token_id u32              = 0
llama_model_loader: - kv  38:                tokenizer.ggml.eos_token_id u32              = 1
llama_model_loader: - kv  39:            tokenizer.ggml.padding_token_id u32              = 128815
llama_model_loader: - kv  40:               tokenizer.ggml.add_bos_token bool             = true
llama_model_loader: - kv  41:               tokenizer.ggml.add_eos_token bool             = false
llama_model_loader: - kv  42:                    tokenizer.chat_template str              = {% if not add_generation_prompt is de...
llama_model_loader: - kv  43:               general.quantization_version u32              = 2
llama_model_loader: - kv  44:                          general.file_type u32              = 15
llama_model_loader: - kv  45:                                   split.no u16              = 0
llama_model_loader: - kv  46:                        split.tensors.count i32              = 1025
llama_model_loader: - kv  47:                                split.count u16              = 0
llama_model_loader: - type  f32:  361 tensors
llama_model_loader: - type q4_K:  606 tensors
llama_model_loader: - type q6_K:   58 tensors
llama_model_load: error loading model: error loading model vocabulary: unknown pre-tokenizer type: 'deepseek-v3'
llama_load_model_from_file: failed to load model
llama_init_from_gpt_params: error: failed to load model '/home/tridu33/workspace/soft/llamafile/models/DeepSeek-R1-Q4_K_M_one/DeepSeek-R1-Q4_K_M.gguf'
main: error: unable to load model

[Tridu33]

i found maybe it called deepseek3 in

llamafile/llama.cpp/llama.cpp

Line 19163 in 29b5f27

} else if (tmpl == "deepseek3" || tmpl_contains(LU8("<｜Assistant｜>")) && tmpl_contains(LU8("<｜User｜>")) && tmpl_contains(LU8("<｜end▁of▁sentence｜>"))) {

, but how can i change pre-tokenizer type from 'deepseek-v3' to the right one?

[Tridu33]

if llamafile support DeepSeek-R1-Q4_K_M.gguf or not?

[Tridu33]

ggml-org/llama.cpp#12021 similar problem here, what i use is the latest main branch of llamafile.

[Tridu33]

https://github.com/ggml-org/llama.cpp/blob/34a846b5847a18d133b360074f1fb485b2632b2d/convert_hf_to_gguf_update.py#L110