main: build = 3051 (5921b8f0) main: built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu main: seed = 1717153808 llama_model_loader: loaded meta data with 25 key-value pairs and 723 tensors from K2-IMat-GGUF/K2.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 = K2 llama_model_loader: - kv 2: llama.block_count u32 = 80 llama_model_loader: - kv 3: llama.context_length u32 = 2048 llama_model_loader: - kv 4: llama.embedding_length u32 = 8192 llama_model_loader: - kv 5: llama.feed_forward_length u32 = 22016 llama_model_loader: - kv 6: llama.attention.head_count u32 = 64 llama_model_loader: - kv 7: llama.attention.head_count_kv u32 = 64 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 = 1 llama_model_loader: - kv 11: llama.vocab_size u32 = 32032 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,32032] = ["", "~~", "~~", "<0x00>", "<... llama_model_loader: - kv 16: tokenizer.ggml.scores arr[f32,32032] = [0.000000, 0.000000, 0.000000, 0.0000... llama_model_loader: - kv 17: tokenizer.ggml.token_type arr[i32,32032] = [2, 3, 3, 6, 6, 6, 6, 6, 6, 6, 6, 6, ... llama_model_loader: - kv 18: tokenizer.ggml.bos_token_id u32 = 2 llama_model_loader: - kv 19: tokenizer.ggml.eos_token_id u32 = 2 llama_model_loader: - kv 20: tokenizer.ggml.unknown_token_id u32 = 0 llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 0 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: general.quantization_version u32 = 2 llama_model_loader: - type f32: 161 tensors llama_model_loader: - type f16: 562 tensors llm_load_vocab: special tokens cache size = 291 llm_load_vocab: token to piece cache size = 0.3373 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 = 32032 llm_load_print_meta: n_merges = 0 llm_load_print_meta: n_ctx_train = 2048 llm_load_print_meta: n_embd = 8192 llm_load_print_meta: n_head = 64 llm_load_print_meta: n_head_kv = 64 llm_load_print_meta: n_layer = 80 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 = 1 llm_load_print_meta: n_embd_k_gqa = 8192 llm_load_print_meta: n_embd_v_gqa = 8192 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 = 22016 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 = 2048 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 = 65B llm_load_print_meta: model ftype = F16 llm_load_print_meta: model params = 65.29 B llm_load_print_meta: model size = 121.61 GiB (16.00 BPW) llm_load_print_meta: general.name = K2 llm_load_print_meta: BOS token = 2 '' llm_load_print_meta: EOS token = 2 '' llm_load_print_meta: UNK token = 0 '' llm_load_print_meta: PAD token = 0 '' llm_load_print_meta: LF token = 13 '<0x0A>' 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 = 0.74 MiB llm_load_tensors: offloading 14 repeating layers to GPU llm_load_tensors: offloaded 14/81 layers to GPU llm_load_tensors: CPU buffer size = 124526.03 MiB llm_load_tensors: CUDA0 buffer size = 21616.88 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 = 1056.00 MiB llama_kv_cache_init: CUDA0 KV buffer size = 224.00 MiB llama_new_context_with_model: KV self size = 1280.00 MiB, K (f16): 640.00 MiB, V (f16): 640.00 MiB llama_new_context_with_model: CUDA_Host output buffer size = 0.12 MiB llama_new_context_with_model: CUDA0 compute buffer size = 579.06 MiB llama_new_context_with_model: CUDA_Host compute buffer size = 33.01 MiB llama_new_context_with_model: graph nodes = 2566 llama_new_context_with_model: graph splits = 730 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 94.683 ms compute_imatrix: computing over 151 chunks with batch_size 512 compute_imatrix: 107.23 seconds per pass - ETA 4 hours 29.85 minutes [1]2.9874,[2]2.5445,[3]2.6271,[4]2.7524,[5]3.1375,[6]3.1526,[7]2.8832,[8]3.2468,[9]3.3474, save_imatrix: stored collected data after 10 chunks in K2-IMat-GGUF/imatrix.dat [10]3.5923,[11]3.7040,[12]3.4774,[13]3.5288,[14]3.7929,[15]4.1212,[16]4.2887,[17]4.5745,[18]4.7454,[19]4.9200, save_imatrix: stored collected data after 20 chunks in K2-IMat-GGUF/imatrix.dat [20]5.0482,[21]5.1606,[22]4.9780,[23]4.7910,[24]4.8409,[25]4.8640,[26]4.8573,[27]4.8312,[28]4.9107,[29]5.0360, save_imatrix: stored collected data after 30 chunks in K2-IMat-GGUF/imatrix.dat [30]5.1757,[31]5.1243,[32]4.9548,[33]4.8161,[34]4.7810,[35]4.7993,[36]4.7667,[37]4.6042,[38]4.5005,[39]4.4803, save_imatrix: stored collected data after 40 chunks in K2-IMat-GGUF/imatrix.dat [40]4.4651,[41]4.4600,[42]4.4838,[43]4.4582,[44]4.4823,[45]4.4822,[46]4.4737,[47]4.5013,[48]4.5763,[49]4.6423, save_imatrix: stored collected data after 50 chunks in K2-IMat-GGUF/imatrix.dat [50]4.6678,[51]4.7174,[52]4.7779,[53]4.8658,[54]4.9522,[55]4.9964,[56]4.9827,[57]4.9548,[58]4.9871,[59]5.0248, save_imatrix: stored collected data after 60 chunks in K2-IMat-GGUF/imatrix.dat [60]5.0988,[61]5.0558,[62]5.0845,[63]5.1063,[64]5.1754,[65]5.2328,[66]5.2571,[67]5.2991,[68]5.3429,[69]5.3598, save_imatrix: stored collected data after 70 chunks in K2-IMat-GGUF/imatrix.dat [70]5.3737,[71]5.3712,[72]5.3795,[73]5.3481,[74]5.2954,[75]5.2986,[76]5.2976,[77]5.3090,[78]5.2743,[79]5.2777, save_imatrix: stored collected data after 80 chunks in K2-IMat-GGUF/imatrix.dat [80]5.2845,[81]5.2587,[82]5.2497,[83]5.2269,[84]5.2329,[85]5.2306,[86]5.2213,[87]5.2190,[88]5.2209,[89]5.1987, save_imatrix: stored collected data after 90 chunks in K2-IMat-GGUF/imatrix.dat [90]5.1873,[91]5.1907,[92]5.1721,[93]5.1488,[94]5.1349,[95]5.0904,[96]5.1054,[97]5.0937,[98]5.0927,[99]5.0769, save_imatrix: stored collected data after 100 chunks in K2-IMat-GGUF/imatrix.dat [100]5.0647,[101]5.0843,[102]5.0563,[103]5.0346,[104]5.0257,[105]5.0408,[106]5.0480,[107]5.0635,[108]5.0796,[109]5.0420, save_imatrix: stored collected data after 110 chunks in K2-IMat-GGUF/imatrix.dat [110]5.0087,[111]4.9742,[112]4.9400,[113]4.9095,[114]4.8773,[115]4.8466,[116]4.8150,[117]4.7943,[118]4.8070,[119]4.8246, save_imatrix: stored collected data after 120 chunks in K2-IMat-GGUF/imatrix.dat [120]4.8683,[121]4.9068,[122]4.9497,[123]4.9869,[124]5.0502,[125]5.1029,[126]5.1172,[127]5.1213,[128]5.0801,[129]5.0814, save_imatrix: stored collected data after 130 chunks in K2-IMat-GGUF/imatrix.dat [130]5.0795,[131]5.0715,[132]5.0527,[133]5.0369,[134]5.0574,[135]5.0784,[136]5.0831,[137]5.0860,[138]5.0985,[139]5.1172, save_imatrix: stored collected data after 140 chunks in K2-IMat-GGUF/imatrix.dat [140]5.1348,[141]5.1434,[142]5.1527,[143]5.1539,[144]5.1406,[145]5.1483,[146]5.1365,[147]5.1246,[148]5.1367,[149]5.1277, save_imatrix: stored collected data after 150 chunks in K2-IMat-GGUF/imatrix.dat [150]5.1280,[151]5.1125, save_imatrix: stored collected data after 151 chunks in K2-IMat-GGUF/imatrix.dat llama_print_timings: load time = 231679.50 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 = 1283171.93 ms / 77312 tokens ( 16.60 ms per token, 60.25 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 = 1410561.69 ms / 77313 tokens Final estimate: PPL = 5.1125 +/- 0.05847