main: build = 3072 (549279d8) main: built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu main: seed = 1717403308 llama_model_loader: loaded meta data with 26 key-value pairs and 363 tensors from dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/dolphin-2.9.2-Phi-3-Medium.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 = llama llama_model_loader: - kv 1: general.name str = dolphin-2.9.2-Phi-3-Medium llama_model_loader: - kv 2: llama.block_count u32 = 40 llama_model_loader: - kv 3: llama.context_length u32 = 4096 llama_model_loader: - kv 4: llama.embedding_length u32 = 5120 llama_model_loader: - kv 5: llama.feed_forward_length u32 = 17920 llama_model_loader: - kv 6: llama.attention.head_count u32 = 40 llama_model_loader: - kv 7: llama.attention.head_count_kv u32 = 10 llama_model_loader: - kv 8: llama.rope.freq_base f32 = 10000.000000 llama_model_loader: - kv 9: llama.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 10: general.file_type u32 = 0 llama_model_loader: - kv 11: llama.vocab_size u32 = 32064 llama_model_loader: - kv 12: llama.rope.dimension_count u32 = 128 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] = [0.000000, 0.000000, 0.000000, 0.0000... llama_model_loader: - kv 17: tokenizer.ggml.token_type arr[i32,32064] = [2, 3, 3, 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 = 32011 llama_model_loader: - kv 20: tokenizer.ggml.unknown_token_id u32 = 0 llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 32009 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 = {% if not add_generation_prompt is de... llama_model_loader: - kv 25: general.quantization_version u32 = 2 llama_model_loader: - type f32: 363 tensors llm_load_vocab: special tokens cache size = 323 llm_load_vocab: token to piece cache size = 0.1687 MB llm_load_print_meta: format = GGUF V3 (latest) llm_load_print_meta: arch = llama 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 = 0 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 = 13B 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 = dolphin-2.9.2-Phi-3-Medium llm_load_print_meta: BOS token = 1 '' llm_load_print_meta: EOS token = 32011 '<|im_end|>' llm_load_print_meta: UNK token = 0 '' llm_load_print_meta: PAD token = 32009 '<|placeholder6|>' 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 = 0.37 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 = 698.88 MiB llama_new_context_with_model: CUDA_Host compute buffer size = 11.01 MiB llama_new_context_with_model: graph nodes = 1286 llama_new_context_with_model: graph splits = 268 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 90.545 ms compute_imatrix: computing over 151 chunks with batch_size 512 compute_imatrix: 2.67 seconds per pass - ETA 6.72 minutes [1]5.4927,[2]4.2476,[3]4.1310,[4]4.5178,[5]4.9728,[6]5.0987,[7]4.5996,[8]5.0199,[9]5.3683, save_imatrix: stored collected data after 10 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [10]5.7273,[11]5.6911,[12]5.3080,[13]5.2015,[14]5.3775,[15]5.7857,[16]5.9135,[17]6.2558,[18]6.3931,[19]6.5133, save_imatrix: stored collected data after 20 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [20]6.6204,[21]6.8296,[22]6.5908,[23]6.3187,[24]6.3938,[25]6.4297,[26]6.3279,[27]6.2584,[28]6.3482,[29]6.5166, save_imatrix: stored collected data after 30 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [30]6.6336,[31]6.5860,[32]6.6940,[33]6.8055,[34]6.9687,[35]6.9646,[36]6.9148,[37]6.6549,[38]6.4872,[39]6.3776, save_imatrix: stored collected data after 40 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [40]6.2646,[41]6.1682,[42]6.1837,[43]6.1322,[44]6.1485,[45]6.1357,[46]6.0911,[47]6.0907,[48]6.1643,[49]6.2343, save_imatrix: stored collected data after 50 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [50]6.2389,[51]6.4028,[52]6.5394,[53]6.7146,[54]6.8627,[55]6.9453,[56]6.8553,[57]6.7499,[58]6.7550,[59]6.8034, save_imatrix: stored collected data after 60 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [60]6.8864,[61]6.7589,[62]6.7750,[63]6.8257,[64]6.9128,[65]6.9681,[66]6.9917,[67]7.0450,[68]7.0888,[69]7.0789, save_imatrix: stored collected data after 70 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [70]7.0734,[71]7.0759,[72]7.1233,[73]7.0709,[74]7.0016,[75]6.9791,[76]7.0247,[77]7.0168,[78]6.9833,[79]6.9746, save_imatrix: stored collected data after 80 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [80]6.9883,[81]6.9484,[82]6.9232,[83]6.8963,[84]6.8935,[85]6.8971,[86]6.8859,[87]6.8902,[88]6.8823,[89]6.8693, save_imatrix: stored collected data after 90 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [90]6.8530,[91]6.8611,[92]6.8256,[93]6.8248,[94]6.7972,[95]6.7539,[96]6.7650,[97]6.7432,[98]6.7502,[99]6.7194, save_imatrix: stored collected data after 100 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [100]6.7120,[101]6.7192,[102]6.6944,[103]6.6651,[104]6.6422,[105]6.6521,[106]6.6467,[107]6.6735,[108]6.6925,[109]6.6445, save_imatrix: stored collected data after 110 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [110]6.5983,[111]6.5500,[112]6.5017,[113]6.4535,[114]6.4125,[115]6.3703,[116]6.3250,[117]6.2976,[118]6.3060,[119]6.3119, save_imatrix: stored collected data after 120 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [120]6.3539,[121]6.3915,[122]6.4386,[123]6.4871,[124]6.5630,[125]6.6257,[126]6.6366,[127]6.6452,[128]6.5921,[129]6.5853, save_imatrix: stored collected data after 130 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [130]6.5508,[131]6.5163,[132]6.4652,[133]6.3948,[134]6.4089,[135]6.4312,[136]6.4288,[137]6.4225,[138]6.4324,[139]6.4425, save_imatrix: stored collected data after 140 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [140]6.4539,[141]6.4556,[142]6.4586,[143]6.4484,[144]6.4286,[145]6.4497,[146]6.4785,[147]6.5129,[148]6.5450,[149]6.5837, save_imatrix: stored collected data after 150 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat [150]6.6157,[151]6.6491, save_imatrix: stored collected data after 151 chunks in dolphin-2.9.2-Phi-3-Medium-IMat-GGUF/imatrix.dat llama_print_timings: load time = 5164.30 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 = 374332.95 ms / 77312 tokens ( 4.84 ms per token, 206.53 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 = 377375.22 ms / 77313 tokens Final estimate: PPL = 6.6491 +/- 0.08775