main: build = 3067 (9422c5e3) main: built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu main: seed = 1717334117 llama_model_loader: loaded meta data with 24 key-value pairs and 387 tensors from Nxcode-CQ-7B-orpo-IMat-GGUF/Nxcode-CQ-7B-orpo.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 = qwen2 llama_model_loader: - kv 1: general.name str = Nxcode-CQ-7B-orpo llama_model_loader: - kv 2: qwen2.block_count u32 = 32 llama_model_loader: - kv 3: qwen2.context_length u32 = 65536 llama_model_loader: - kv 4: qwen2.embedding_length u32 = 4096 llama_model_loader: - kv 5: qwen2.feed_forward_length u32 = 13440 llama_model_loader: - kv 6: qwen2.attention.head_count u32 = 32 llama_model_loader: - kv 7: qwen2.attention.head_count_kv u32 = 4 llama_model_loader: - kv 8: qwen2.rope.freq_base f32 = 1000000.000000 llama_model_loader: - kv 9: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 10: general.file_type u32 = 0 llama_model_loader: - kv 11: tokenizer.ggml.model str = llama llama_model_loader: - kv 12: tokenizer.ggml.pre str = default llama_model_loader: - kv 13: tokenizer.ggml.tokens arr[str,92416] = ["", "", "<|endoftext|>", "<|... llama_model_loader: - kv 14: tokenizer.ggml.scores arr[f32,92416] = [0.000000, 0.000000, 0.000000, 0.0000... llama_model_loader: - kv 15: tokenizer.ggml.token_type arr[i32,92416] = [2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, ... llama_model_loader: - kv 16: tokenizer.ggml.bos_token_id u32 = 2 llama_model_loader: - kv 17: tokenizer.ggml.eos_token_id u32 = 4 llama_model_loader: - kv 18: tokenizer.ggml.unknown_token_id u32 = 0 llama_model_loader: - kv 19: tokenizer.ggml.padding_token_id u32 = 92298 llama_model_loader: - kv 20: tokenizer.ggml.add_bos_token bool = false llama_model_loader: - kv 21: tokenizer.ggml.add_eos_token bool = false llama_model_loader: - kv 22: tokenizer.chat_template str = {% for message in messages %}{% if lo... llama_model_loader: - kv 23: general.quantization_version u32 = 2 llama_model_loader: - type f32: 387 tensors llm_load_vocab: special tokens cache size = 407 llm_load_vocab: token to piece cache size = 0.9943 MB llm_load_print_meta: format = GGUF V3 (latest) llm_load_print_meta: arch = qwen2 llm_load_print_meta: vocab type = SPM llm_load_print_meta: n_vocab = 92416 llm_load_print_meta: n_merges = 0 llm_load_print_meta: n_ctx_train = 65536 llm_load_print_meta: n_embd = 4096 llm_load_print_meta: n_head = 32 llm_load_print_meta: n_head_kv = 4 llm_load_print_meta: n_layer = 32 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 = 8 llm_load_print_meta: n_embd_k_gqa = 512 llm_load_print_meta: n_embd_v_gqa = 512 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 = 13440 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 = 1000000.0 llm_load_print_meta: freq_scale_train = 1 llm_load_print_meta: n_yarn_orig_ctx = 65536 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 = 7B llm_load_print_meta: model ftype = all F32 llm_load_print_meta: model params = 7.25 B llm_load_print_meta: model size = 27.01 GiB (32.00 BPW) llm_load_print_meta: general.name = Nxcode-CQ-7B-orpo llm_load_print_meta: BOS token = 2 '<|endoftext|>' llm_load_print_meta: EOS token = 4 '<|im_end|>' llm_load_print_meta: UNK token = 0 '' llm_load_print_meta: PAD token = 92298 '' llm_load_print_meta: LF token = 1396 '<0x0A>' llm_load_print_meta: EOT token = 2 '<|endoftext|>' 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 21 repeating layers to GPU llm_load_tensors: offloaded 21/33 layers to GPU llm_load_tensors: CPU buffer size = 27657.64 MiB llm_load_tensors: CUDA0 buffer size = 16255.07 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 = 1000000.0 llama_new_context_with_model: freq_scale = 1 llama_kv_cache_init: CUDA_Host KV buffer size = 11.00 MiB llama_kv_cache_init: CUDA0 KV buffer size = 21.00 MiB llama_new_context_with_model: KV self size = 32.00 MiB, K (f16): 16.00 MiB, V (f16): 16.00 MiB llama_new_context_with_model: CUDA_Host output buffer size = 0.35 MiB llama_new_context_with_model: CUDA0 compute buffer size = 1632.50 MiB llama_new_context_with_model: CUDA_Host compute buffer size = 9.01 MiB llama_new_context_with_model: graph nodes = 1126 llama_new_context_with_model: graph splits = 158 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 98.978 ms compute_imatrix: computing over 138 chunks with batch_size 512 compute_imatrix: 1.10 seconds per pass - ETA 2.52 minutes [1]10.0166,[2]7.7808,[3]7.8168,[4]8.8446,[5]8.6711,[6]9.1749,[7]10.6247,[8]9.9414,[9]11.3947, save_imatrix: stored collected data after 10 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [10]11.5342,[11]9.8949,[12]10.7034,[13]11.7077,[14]12.1832,[15]12.2311,[16]12.5750,[17]12.8830,[18]13.2199,[19]13.6752, save_imatrix: stored collected data after 20 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [20]12.6298,[21]12.3094,[22]12.6268,[23]12.7491,[24]12.2356,[25]12.2872,[26]11.9565,[27]12.1363,[28]12.4309,[29]12.5358, save_imatrix: stored collected data after 30 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [30]12.7107,[31]13.1703,[32]13.2835,[33]13.2356,[34]12.3675,[35]11.6135,[36]11.4960,[37]11.3605,[38]11.1926,[39]11.2979, save_imatrix: stored collected data after 40 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [40]11.6469,[41]11.6846,[42]11.7268,[43]11.8902,[44]12.0713,[45]12.4834,[46]12.8479,[47]13.2627,[48]13.6070,[49]13.8156, save_imatrix: stored collected data after 50 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [50]13.7637,[51]13.6509,[52]13.7997,[53]13.9016,[54]13.9780,[55]13.7988,[56]13.5633,[57]13.7076,[58]13.8893,[59]14.0737, save_imatrix: stored collected data after 60 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [60]14.1681,[61]14.2382,[62]14.2079,[63]14.2551,[64]14.0811,[65]13.8605,[66]13.8274,[67]13.7997,[68]13.8085,[69]13.8360, save_imatrix: stored collected data after 70 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [70]13.8269,[71]13.8219,[72]13.8300,[73]13.7448,[74]13.6263,[75]13.6240,[76]13.5531,[77]13.4955,[78]13.3634,[79]13.3624, save_imatrix: stored collected data after 80 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [80]13.3119,[81]13.1828,[82]13.1783,[83]13.0645,[84]12.9864,[85]12.8800,[86]12.8612,[87]12.9561,[88]12.9683,[89]12.9217, save_imatrix: stored collected data after 90 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [90]12.8961,[91]12.9203,[92]12.6766,[93]12.6844,[94]12.7573,[95]12.7819,[96]12.8019,[97]12.7789,[98]12.5869,[99]12.3970, save_imatrix: stored collected data after 100 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [100]12.2213,[101]12.0474,[102]11.8728,[103]11.7025,[104]11.5473,[105]11.3917,[106]11.3534,[107]11.3443,[108]11.3572,[109]11.4759, save_imatrix: stored collected data after 110 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [110]11.5911,[111]11.7258,[112]11.8873,[113]12.0252,[114]12.0364,[115]12.0735,[116]11.9485,[117]11.9701,[118]11.9270,[119]11.8552, save_imatrix: stored collected data after 120 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [120]11.7679,[121]11.7419,[122]11.7789,[123]11.7956,[124]11.8587,[125]11.8861,[126]11.9464,[127]11.9999,[128]12.0050,[129]12.0363, save_imatrix: stored collected data after 130 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat [130]12.0133,[131]11.9473,[132]12.0626,[133]12.1655,[134]12.2616,[135]12.3497,[136]12.4361,[137]12.5321,[138]12.5893, save_imatrix: stored collected data after 138 chunks in Nxcode-CQ-7B-orpo-IMat-GGUF/imatrix.dat llama_print_timings: load time = 2774.51 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 = 137069.11 ms / 70656 tokens ( 1.94 ms per token, 515.48 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 = 139467.01 ms / 70657 tokens Final estimate: PPL = 12.5893 +/- 0.19777