File size: 12,120 Bytes
94e9f67 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 |
main: build = 3008 (1d8fca72)
main: built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu
main: seed = 1716817514
llama_model_loader: loaded meta data with 22 key-value pairs and 219 tensors from internlm2-math-plus-1_8b-IMat-GGUF/internlm2-math-plus-1_8b.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 = internlm2
llama_model_loader: - kv 1: general.name str = InternLM2
llama_model_loader: - kv 2: internlm2.context_length u32 = 8192
llama_model_loader: - kv 3: internlm2.block_count u32 = 24
llama_model_loader: - kv 4: internlm2.embedding_length u32 = 2048
llama_model_loader: - kv 5: internlm2.feed_forward_length u32 = 8192
llama_model_loader: - kv 6: internlm2.rope.freq_base f32 = 1000000.000000
llama_model_loader: - kv 7: internlm2.attention.head_count u32 = 16
llama_model_loader: - kv 8: internlm2.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 9: internlm2.attention.head_count_kv u32 = 8
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,92544] = ["<unk>", "<s>", "</s>", "<0x00>", "<...
llama_model_loader: - kv 14: tokenizer.ggml.scores arr[f32,92544] = [0.000000, 0.000000, 0.000000, 0.0000...
llama_model_loader: - kv 15: tokenizer.ggml.token_type arr[i32,92544] = [2, 3, 3, 6, 6, 6, 6, 6, 6, 6, 6, 6, ...
llama_model_loader: - kv 16: tokenizer.ggml.add_space_prefix bool = false
llama_model_loader: - kv 17: tokenizer.ggml.bos_token_id u32 = 1
llama_model_loader: - kv 18: tokenizer.ggml.eos_token_id u32 = 2
llama_model_loader: - kv 19: tokenizer.ggml.padding_token_id u32 = 2
llama_model_loader: - kv 20: tokenizer.chat_template str = {{ bos_token }}{% for message in mess...
llama_model_loader: - kv 21: general.quantization_version u32 = 2
llama_model_loader: - type f32: 219 tensors
llm_load_vocab: mismatch in special tokens definition ( 405/92544 vs 259/92544 ).
llm_load_print_meta: format = GGUF V3 (latest)
llm_load_print_meta: arch = internlm2
llm_load_print_meta: vocab type = SPM
llm_load_print_meta: n_vocab = 92544
llm_load_print_meta: n_merges = 0
llm_load_print_meta: n_ctx_train = 8192
llm_load_print_meta: n_embd = 2048
llm_load_print_meta: n_head = 16
llm_load_print_meta: n_head_kv = 8
llm_load_print_meta: n_layer = 24
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 = 2
llm_load_print_meta: n_embd_k_gqa = 1024
llm_load_print_meta: n_embd_v_gqa = 1024
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 = 8192
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 = 1000000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_yarn_orig_ctx = 8192
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 = ?B
llm_load_print_meta: model ftype = all F32
llm_load_print_meta: model params = 1.89 B
llm_load_print_meta: model size = 7.04 GiB (32.00 BPW)
llm_load_print_meta: general.name = InternLM2
llm_load_print_meta: BOS token = 1 '<s>'
llm_load_print_meta: EOS token = 2 '</s>'
llm_load_print_meta: UNK token = 0 '<unk>'
llm_load_print_meta: PAD token = 2 '</s>'
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 = 0.22 MiB
llm_load_tensors: offloading 24 repeating layers to GPU
llm_load_tensors: offloading non-repeating layers to GPU
llm_load_tensors: offloaded 25/25 layers to GPU
llm_load_tensors: CPU buffer size = 723.00 MiB
llm_load_tensors: CUDA0 buffer size = 6483.38 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: CUDA0 KV buffer size = 48.00 MiB
llama_new_context_with_model: KV self size = 48.00 MiB, K (f16): 24.00 MiB, V (f16): 24.00 MiB
llama_new_context_with_model: CUDA_Host output buffer size = 0.35 MiB
llama_new_context_with_model: CUDA0 compute buffer size = 184.75 MiB
llama_new_context_with_model: CUDA_Host compute buffer size = 5.01 MiB
llama_new_context_with_model: graph nodes = 774
llama_new_context_with_model: graph splits = 2
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 149.089 ms
compute_imatrix: computing over 209 chunks with batch_size 512
compute_imatrix: 0.25 seconds per pass - ETA 0.87 minutes
[1]11.1312,[2]8.6363,[3]7.5192,[4]8.7685,[5]8.6457,[6]8.0250,[7]9.1250,[8]9.1274,[9]9.9765,
save_imatrix: stored collected data after 10 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[10]10.3272,[11]11.1727,[12]11.3352,[13]12.9598,[14]13.4077,[15]14.4618,[16]15.1761,[17]15.8490,[18]14.8805,[19]15.3401,
save_imatrix: stored collected data after 20 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[20]15.2807,[21]14.5097,[22]14.4950,[23]13.4617,[24]13.0109,[25]12.2200,[26]12.2099,[27]12.7869,[28]12.7652,[29]13.2829,
save_imatrix: stored collected data after 30 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[30]13.7325,[31]13.6520,[32]12.9249,[33]12.4811,[34]12.2812,[35]12.2370,[36]12.0979,[37]12.4393,[38]12.8247,[39]13.0730,
save_imatrix: stored collected data after 40 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[40]13.4472,[41]13.7132,[42]14.1578,[43]14.5769,[44]15.0185,[45]15.1708,[46]15.2992,[47]15.2611,[48]15.1091,[49]15.4024,
save_imatrix: stored collected data after 50 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[50]15.7380,[51]15.8763,[52]16.2449,[53]16.2822,[54]16.5116,[55]16.7323,[56]17.0042,[57]17.1444,[58]17.3613,[59]17.3927,
save_imatrix: stored collected data after 60 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[60]17.3337,[61]17.6352,[62]17.9523,[63]18.4370,[64]18.3717,[65]18.1018,[66]17.8581,[67]17.6437,[68]17.4301,[69]17.1963,
save_imatrix: stored collected data after 70 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[70]17.0333,[71]16.9842,[72]16.7208,[73]16.3945,[74]16.5737,[75]16.7233,[76]16.7770,[77]16.7379,[78]16.9561,[79]17.0106,
save_imatrix: stored collected data after 80 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[80]17.1238,[81]17.1064,[82]17.1041,[83]17.2104,[84]17.2325,[85]17.2776,[86]17.2832,[87]17.3172,[88]17.2650,[89]17.3473,
save_imatrix: stored collected data after 90 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[90]17.4284,[91]17.5131,[92]17.4812,[93]17.3928,[94]17.2928,[95]17.2508,[96]17.1120,[97]17.1103,[98]17.0537,[99]16.9868,
save_imatrix: stored collected data after 100 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[100]16.8327,[101]16.8049,[102]16.7018,[103]16.5576,[104]16.4292,[105]16.3660,[106]16.2770,[107]16.1450,[108]16.0753,[109]16.0705,
save_imatrix: stored collected data after 110 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[110]16.0731,[111]15.9900,[112]15.9993,[113]16.0081,[114]15.9285,[115]15.8458,[116]15.9694,[117]15.9543,[118]15.9839,[119]15.8023,
save_imatrix: stored collected data after 120 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[120]15.6437,[121]15.4694,[122]15.2755,[123]15.1011,[124]14.9437,[125]14.7883,[126]14.7195,[127]14.6397,[128]14.5264,[129]14.4086,
save_imatrix: stored collected data after 130 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[130]14.3366,[131]14.2389,[132]14.1439,[133]14.0853,[134]13.9819,[135]13.8918,[136]13.8522,[137]13.7711,[138]13.6823,[139]13.6393,
save_imatrix: stored collected data after 140 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[140]13.5538,[141]13.4733,[142]13.5984,[143]13.8155,[144]14.0751,[145]14.2823,[146]14.3138,[147]14.3540,[148]14.4225,[149]14.5055,
save_imatrix: stored collected data after 150 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[150]14.5680,[151]14.5877,[152]14.6063,[153]14.7027,[154]14.7721,[155]14.8630,[156]14.8986,[157]14.9890,[158]15.0815,[159]15.1157,
save_imatrix: stored collected data after 160 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[160]15.1994,[161]15.2529,[162]15.2996,[163]15.3659,[164]15.4310,[165]15.4594,[166]15.5148,[167]15.5710,[168]15.6024,[169]15.6428,
save_imatrix: stored collected data after 170 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[170]15.6873,[171]15.7017,[172]15.7554,[173]15.8180,[174]15.8137,[175]15.9375,[176]16.0642,[177]16.2008,[178]16.3725,[179]16.4786,
save_imatrix: stored collected data after 180 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[180]16.5336,[181]16.4838,[182]16.5160,[183]16.5718,[184]16.6541,[185]16.6811,[186]16.6915,[187]16.7102,[188]16.7611,[189]16.7740,
save_imatrix: stored collected data after 190 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[190]16.7763,[191]16.8201,[192]16.8665,[193]16.9179,[194]16.8887,[195]16.9311,[196]16.9176,[197]16.9523,[198]16.9722,[199]17.0944,
save_imatrix: stored collected data after 200 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
[200]16.9780,[201]17.0414,[202]17.0243,[203]17.1840,[204]17.3468,[205]17.4907,[206]17.6111,[207]17.7104,[208]17.6297,[209]17.5631,
save_imatrix: stored collected data after 209 chunks in internlm2-math-plus-1_8b-IMat-GGUF/imatrix.dat
llama_print_timings: load time = 1311.79 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 = 32872.80 ms / 107008 tokens ( 0.31 ms per token, 3255.21 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 = 35006.68 ms / 107009 tokens
Final estimate: PPL = 17.5631 +/- 0.23703
|