Eval bug: Error running Phi4-mini gguf: unknown pre-tokenizer type: 'gpt-4o'

[crisdesivo]

Name and Version

version: 3400 (97bdd26)
built with cc (Ubuntu 9.4.0-1ubuntu1~20.04.2) 9.4.0 for x86_64-linux-gnu

Operating systems

Linux

GGML backends

CUDA

Hardware

NVIDIA GeForce RTX 3050

Models

microsoft_Phi-4-mini-instruct-Q8_0.gguf

Problem description & steps to reproduce

When I run the llama server with the phi4 mini gguf I get the error

./llama-server -m microsoft_Phi-4-mini-instruct-Q8_0.gguf

First Bad Commit

No response

Relevant log output

INFO [                    main] build info | tid="139955235852288" timestamp=1740794211 build=3400 commit="97bdd26e"
INFO [                    main] system info | tid="139955235852288" timestamp=1740794211 n_threads=4 n_threads_batch=-1 total_threads=8 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 1 | AVX512 = 1 | AVX512_VBMI = 1 | AVX512_VNNI = 1 | AVX512_BF16 = 0 | FMA = 1 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | "
llama_model_loader: loaded meta data with 40 key-value pairs and 196 tensors from microsoft_Phi-4-mini-instruct-Q8_0.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              = phi3
llama_model_loader: - kv   1:              phi3.rope.scaling.attn_factor f32              = 1.190238
llama_model_loader: - kv   2:                               general.type str              = model
llama_model_loader: - kv   3:                               general.name str              = Phi 4 Mini Instruct
llama_model_loader: - kv   4:                           general.finetune str              = instruct
llama_model_loader: - kv   5:                           general.basename str              = Phi-4
llama_model_loader: - kv   6:                         general.size_label str              = mini
llama_model_loader: - kv   7:                            general.license str              = mit
llama_model_loader: - kv   8:                       general.license.link str              = https://huggingface.co/microsoft/Phi-...
llama_model_loader: - kv   9:                               general.tags arr[str,3]       = ["nlp", "code", "text-generation"]
llama_model_loader: - kv  10:                          general.languages arr[str,1]       = ["multilingual"]
llama_model_loader: - kv  11:                        phi3.context_length u32              = 131072
llama_model_loader: - kv  12:  phi3.rope.scaling.original_context_length u32              = 4096
llama_model_loader: - kv  13:                      phi3.embedding_length u32              = 3072
llama_model_loader: - kv  14:                   phi3.feed_forward_length u32              = 8192
llama_model_loader: - kv  15:                           phi3.block_count u32              = 32
llama_model_loader: - kv  16:                  phi3.attention.head_count u32              = 24
llama_model_loader: - kv  17:               phi3.attention.head_count_kv u32              = 8
llama_model_loader: - kv  18:      phi3.attention.layer_norm_rms_epsilon f32              = 0.000010
llama_model_loader: - kv  19:                  phi3.rope.dimension_count u32              = 96
llama_model_loader: - kv  20:                        phi3.rope.freq_base f32              = 10000.000000
llama_model_loader: - kv  21:              phi3.attention.sliding_window u32              = 262144
llama_model_loader: - kv  22:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  23:                         tokenizer.ggml.pre str              = gpt-4o
llama_model_loader: - kv  24:                      tokenizer.ggml.tokens arr[str,200064]  = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv  25:                  tokenizer.ggml.token_type arr[i32,200064]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv  26:                      tokenizer.ggml.merges arr[str,199742]  = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "e r", ...
llama_model_loader: - kv  27:                tokenizer.ggml.bos_token_id u32              = 199999
llama_model_loader: - kv  28:                tokenizer.ggml.eos_token_id u32              = 199999
llama_model_loader: - kv  29:            tokenizer.ggml.unknown_token_id u32              = 199999
llama_model_loader: - kv  30:            tokenizer.ggml.padding_token_id u32              = 199999
llama_model_loader: - kv  31:               tokenizer.ggml.add_bos_token bool             = false
llama_model_loader: - kv  32:               tokenizer.ggml.add_eos_token bool             = false
llama_model_loader: - kv  33:                    tokenizer.chat_template str              = {% for message in messages %}{% if me...
llama_model_loader: - kv  34:               general.quantization_version u32              = 2
llama_model_loader: - kv  35:                          general.file_type u32              = 7
llama_model_loader: - kv  36:                      quantize.imatrix.file str              = /models_out/Phi-4-mini-instruct-GGUF/...
llama_model_loader: - kv  37:                   quantize.imatrix.dataset str              = /training_dir/calibration_datav3.txt
llama_model_loader: - kv  38:             quantize.imatrix.entries_count i32              = 128
llama_model_loader: - kv  39:              quantize.imatrix.chunks_count i32              = 123
llama_model_loader: - type  f32:   67 tensors
llama_model_loader: - type q8_0:  129 tensors
llama_model_load: error loading model: error loading model vocabulary: unknown pre-tokenizer type: 'gpt-4o'
llama_load_model_from_file: failed to load model
llama_init_from_gpt_params: error: failed to load model 'microsoft_Phi-4-mini-instruct-Q8_0.gguf'
 ERR [              load_model] unable to load model | tid="139955235852288" timestamp=1740794211 model="microsoft_Phi-4-mini-instruct-Q8_0.gguf"
munmap_chunk(): invalid pointer
Aborted (core dumped)

[vlamirand]

b4792

[crisdesivo]

I

b4792

It was a versions issue, however I had pulled the latest code.

The problem was that since the last version I had installed, the location of the llama-cli file was changed (it used to be on the root directory, now it's on build/bin), so it didn't overwrite the one I was using to check the version or to do the actual inference.

Thank you, hope this helps someone else so I don't feel that dumb,