main: build = 2998 (9588f196) main: built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu main: seed = 1716718166 llama_model_loader: loaded meta data with 26 key-value pairs and 243 tensors from Phi-3-medium-4k-instruct-IMat-GGUF/Phi-3-medium-4k-instruct.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: general.name str = Phi3 llama_model_loader: - kv 2: phi3.context_length u32 = 4096 llama_model_loader: - kv 3: phi3.rope.scaling.original_context_length u32 = 4096 llama_model_loader: - kv 4: phi3.embedding_length u32 = 5120 llama_model_loader: - kv 5: phi3.feed_forward_length u32 = 17920 llama_model_loader: - kv 6: phi3.block_count u32 = 40 llama_model_loader: - kv 7: phi3.attention.head_count u32 = 40 llama_model_loader: - kv 8: phi3.attention.head_count_kv u32 = 10 llama_model_loader: - kv 9: phi3.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 10: phi3.rope.dimension_count u32 = 128 llama_model_loader: - kv 11: phi3.rope.freq_base f32 = 10000.000000 llama_model_loader: - kv 12: general.file_type u32 = 0 llama_model_loader: - kv 13: tokenizer.ggml.model str = llama llama_model_loader: - kv 14: tokenizer.ggml.pre str = default llama_model_loader: - kv 15: tokenizer.ggml.tokens arr[str,32064] = ["", "~~", "~~", "<0x00>", "<... llama_model_loader: - kv 16: tokenizer.ggml.scores arr[f32,32064] = [-1000.000000, -1000.000000, -1000.00... llama_model_loader: - kv 17: tokenizer.ggml.token_type arr[i32,32064] = [3, 3, 4, 6, 6, 6, 6, 6, 6, 6, 6, 6, ... llama_model_loader: - kv 18: tokenizer.ggml.bos_token_id u32 = 1 llama_model_loader: - kv 19: tokenizer.ggml.eos_token_id u32 = 32000 llama_model_loader: - kv 20: tokenizer.ggml.unknown_token_id u32 = 0 llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 32000 llama_model_loader: - kv 22: tokenizer.ggml.add_bos_token bool = false llama_model_loader: - kv 23: tokenizer.ggml.add_eos_token bool = false llama_model_loader: - kv 24: tokenizer.chat_template str = {% for message in messages %}{% if (m... llama_model_loader: - kv 25: general.quantization_version u32 = 2 llama_model_loader: - type f32: 243 tensors llm_load_vocab: special tokens definition check successful ( 323/32064 ). llm_load_print_meta: format = GGUF V3 (latest) llm_load_print_meta: arch = phi3 llm_load_print_meta: vocab type = SPM llm_load_print_meta: n_vocab = 32064 llm_load_print_meta: n_merges = 0 llm_load_print_meta: n_ctx_train = 4096 llm_load_print_meta: n_embd = 5120 llm_load_print_meta: n_head = 40 llm_load_print_meta: n_head_kv = 10 llm_load_print_meta: n_layer = 40 llm_load_print_meta: n_rot = 128 llm_load_print_meta: n_embd_head_k = 128 llm_load_print_meta: n_embd_head_v = 128 llm_load_print_meta: n_gqa = 4 llm_load_print_meta: n_embd_k_gqa = 1280 llm_load_print_meta: n_embd_v_gqa = 1280 llm_load_print_meta: f_norm_eps = 0.0e+00 llm_load_print_meta: f_norm_rms_eps = 1.0e-05 llm_load_print_meta: f_clamp_kqv = 0.0e+00 llm_load_print_meta: f_max_alibi_bias = 0.0e+00 llm_load_print_meta: f_logit_scale = 0.0e+00 llm_load_print_meta: n_ff = 17920 llm_load_print_meta: n_expert = 0 llm_load_print_meta: n_expert_used = 0 llm_load_print_meta: causal attn = 1 llm_load_print_meta: pooling type = 0 llm_load_print_meta: rope type = 2 llm_load_print_meta: rope scaling = linear llm_load_print_meta: freq_base_train = 10000.0 llm_load_print_meta: freq_scale_train = 1 llm_load_print_meta: n_yarn_orig_ctx = 4096 llm_load_print_meta: rope_finetuned = unknown llm_load_print_meta: ssm_d_conv = 0 llm_load_print_meta: ssm_d_inner = 0 llm_load_print_meta: ssm_d_state = 0 llm_load_print_meta: ssm_dt_rank = 0 llm_load_print_meta: model type = 14B llm_load_print_meta: model ftype = all F32 llm_load_print_meta: model params = 13.96 B llm_load_print_meta: model size = 52.01 GiB (32.00 BPW) llm_load_print_meta: general.name = Phi3 llm_load_print_meta: BOS token = 1 '' llm_load_print_meta: EOS token = 32000 '<|endoftext|>' llm_load_print_meta: UNK token = 0 '' llm_load_print_meta: PAD token = 32000 '<|endoftext|>' llm_load_print_meta: LF token = 13 '<0x0A>' llm_load_print_meta: EOT token = 32007 '<|end|>' ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no ggml_cuda_init: CUDA_USE_TENSOR_CORES: yes ggml_cuda_init: found 1 CUDA devices: Device 0: NVIDIA GeForce RTX 4090, compute capability 8.9, VMM: yes llm_load_tensors: ggml ctx size = 0.28 MiB llm_load_tensors: offloading 16 repeating layers to GPU llm_load_tensors: offloaded 16/41 layers to GPU llm_load_tensors: CPU buffer size = 53254.08 MiB llm_load_tensors: CUDA0 buffer size = 20800.62 MiB ....................................................................................... llama_new_context_with_model: n_ctx = 512 llama_new_context_with_model: n_batch = 512 llama_new_context_with_model: n_ubatch = 512 llama_new_context_with_model: flash_attn = 0 llama_new_context_with_model: freq_base = 10000.0 llama_new_context_with_model: freq_scale = 1 llama_kv_cache_init: CUDA_Host KV buffer size = 60.00 MiB llama_kv_cache_init: CUDA0 KV buffer size = 40.00 MiB llama_new_context_with_model: KV self size = 100.00 MiB, K (f16): 50.00 MiB, V (f16): 50.00 MiB llama_new_context_with_model: CUDA_Host output buffer size = 0.12 MiB llama_new_context_with_model: CUDA0 compute buffer size = 870.00 MiB llama_new_context_with_model: CUDA_Host compute buffer size = 11.01 MiB llama_new_context_with_model: graph nodes = 1606 llama_new_context_with_model: graph splits = 196 system_info: n_threads = 25 / 32 | AVX = 1 | AVX_VNNI = 0 | AVX2 = 1 | AVX512 = 1 | AVX512_VBMI = 1 | AVX512_VNNI = 1 | AVX512_BF16 = 1 | 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 | compute_imatrix: tokenizing the input .. compute_imatrix: tokenization took 135.895 ms compute_imatrix: computing over 234 chunks with batch_size 512 compute_imatrix: 2.36 seconds per pass - ETA 9.20 minutes [1]4.4628,[2]3.7699,[3]3.5932,[4]3.9198,[5]4.3289,[6]4.4368,[7]3.9836,[8]4.3653,[9]4.6585, save_imatrix: stored collected data after 10 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [10]4.9541,[11]4.9480,[12]4.6704,[13]4.7845,[14]4.6887,[15]5.0783,[16]5.1954,[17]5.4749,[18]5.6022,[19]5.7812, save_imatrix: stored collected data after 20 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [20]5.9171,[21]5.9895,[22]6.1795,[23]5.9463,[24]5.8125,[25]5.8054,[26]5.5149,[27]5.2993,[28]5.0345,[29]5.0141, save_imatrix: stored collected data after 30 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [30]5.1213,[31]5.1730,[32]5.2166,[33]5.2044,[34]5.2274,[35]5.2328,[36]5.0543,[37]4.9445,[38]4.8950,[39]4.8788, save_imatrix: stored collected data after 40 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [40]4.8606,[41]4.7968,[42]4.8311,[43]4.8610,[44]4.8978,[45]4.9598,[46]5.0218,[47]5.0940,[48]5.2139,[49]5.3086, save_imatrix: stored collected data after 50 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [50]5.4040,[51]5.4848,[52]5.5642,[53]5.5378,[54]5.4705,[55]5.4204,[56]5.4960,[57]5.5296,[58]5.5430,[59]5.5992, save_imatrix: stored collected data after 60 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [60]5.6675,[61]5.6860,[62]5.7221,[63]5.7471,[64]5.7925,[65]5.8159,[66]5.8408,[67]5.8709,[68]5.9073,[69]5.9693, save_imatrix: stored collected data after 70 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [70]6.0047,[71]6.0360,[72]6.0659,[73]6.0194,[74]5.9749,[75]5.8743,[76]5.7833,[77]5.7353,[78]5.6501,[79]5.5922, save_imatrix: stored collected data after 80 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [80]5.5283,[81]5.4407,[82]5.3750,[83]5.3324,[84]5.3255,[85]5.3516,[86]5.3683,[87]5.3982,[88]5.4234,[89]5.4060, save_imatrix: stored collected data after 90 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [90]5.3353,[91]5.3312,[92]5.3133,[93]5.3176,[94]5.3154,[95]5.3198,[96]5.3288,[97]5.3315,[98]5.3151,[99]5.2879, save_imatrix: stored collected data after 100 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [100]5.3000,[101]5.3195,[102]5.3191,[103]5.3003,[104]5.2605,[105]5.2504,[106]5.2643,[107]5.2818,[108]5.2678,[109]5.2625, save_imatrix: stored collected data after 110 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [110]5.2506,[111]5.2598,[112]5.2716,[113]5.2719,[114]5.2876,[115]5.2925,[116]5.2903,[117]5.2905,[118]5.3000,[119]5.2865, save_imatrix: stored collected data after 120 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [120]5.2948,[121]5.2863,[122]5.2687,[123]5.2847,[124]5.2808,[125]5.2896,[126]5.2775,[127]5.2796,[128]5.2894,[129]5.2756, save_imatrix: stored collected data after 130 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [130]5.2624,[131]5.2530,[132]5.2541,[133]5.2180,[134]5.2169,[135]5.2011,[136]5.1883,[137]5.1714,[138]5.1505,[139]5.1312, save_imatrix: stored collected data after 140 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [140]5.1140,[141]5.0956,[142]5.0753,[143]5.0715,[144]5.0631,[145]5.0442,[146]5.0218,[147]5.0165,[148]4.9987,[149]4.9886, save_imatrix: stored collected data after 150 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [150]4.9740,[151]4.9623,[152]4.9518,[153]4.9354,[154]4.9237,[155]4.9383,[156]4.9159,[157]4.9089,[158]4.9125,[159]4.9078, save_imatrix: stored collected data after 160 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [160]4.9054,[161]4.9050,[162]4.9035,[163]4.9103,[164]4.9146,[165]4.9272,[166]4.9297,[167]4.9279,[168]4.9327,[169]4.9389, save_imatrix: stored collected data after 170 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [170]4.9515,[171]4.9474,[172]4.9533,[173]4.9687,[174]4.9736,[175]4.9902,[176]5.0002,[177]5.0093,[178]5.0167,[179]5.0399, save_imatrix: stored collected data after 180 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [180]5.0496,[181]5.0898,[182]5.1043,[183]5.1277,[184]5.1366,[185]5.1447,[186]5.1553,[187]5.1610,[188]5.1536,[189]5.1587, save_imatrix: stored collected data after 190 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [190]5.1661,[191]5.1774,[192]5.1832,[193]5.2059,[194]5.1981,[195]5.1749,[196]5.2077,[197]5.2368,[198]5.2636,[199]5.3059, save_imatrix: stored collected data after 200 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [200]5.3418,[201]5.3512,[202]5.3572,[203]5.3285,[204]5.3301,[205]5.3366,[206]5.3587,[207]5.3574,[208]5.3609,[209]5.3638, save_imatrix: stored collected data after 210 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [210]5.3727,[211]5.3853,[212]5.3860,[213]5.3846,[214]5.3918,[215]5.4068,[216]5.4228,[217]5.4267,[218]5.4235,[219]5.4214, save_imatrix: stored collected data after 220 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [220]5.4164,[221]5.4167,[222]5.4165,[223]5.4248,[224]5.4093,[225]5.4128,[226]5.4006,[227]5.4274,[228]5.4536,[229]5.4882, save_imatrix: stored collected data after 230 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat [230]5.5176,[231]5.5313,[232]5.5115,[233]5.4901,[234]5.4639, save_imatrix: stored collected data after 234 chunks in Phi-3-medium-4k-instruct-IMat-GGUF/imatrix.dat llama_print_timings: load time = 5026.25 ms llama_print_timings: sample time = 0.00 ms / 1 runs ( 0.00 ms per token, inf tokens per second) llama_print_timings: prompt eval time = 528340.36 ms / 119808 tokens ( 4.41 ms per token, 226.76 tokens per second) llama_print_timings: eval time = 0.00 ms / 1 runs ( 0.00 ms per token, inf tokens per second) llama_print_timings: total time = 531821.72 ms / 119809 tokens Final estimate: PPL = 5.4639 +/- 0.05304