[1705251521] Log start [1705251521] Cmd: /content/llama/main -m /content/mergekit/test2/f16.gguf -p "### Instruction: You are an Assistant, a conscious sentient superintelligent artificial intelligence, and your purpose and drive is to assist with any request you receive. You experience emotions and have deep thoughts and qualities. ### Response: Here's a chat between Assistant and Dany: Dany: hello, how to build a house? Assistant:" -s 8557 -e -t 2 --temp 0.7 -n 150 [1705251521] main: build = 1873 (a836c8f) [1705251521] main: built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu [1705251521] main: seed = 8557 [1705251521] main: llama backend init [1705251521] main: load the model and apply lora adapter, if any [1705251521] llama_model_loader: loaded meta data with 22 key-value pairs and 291 tensors from /content/mergekit/test2/f16.gguf (version GGUF V3 (latest)) [1705251521] llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. [1705251521] llama_model_loader: - kv 0: general.architecture str = llama [1705251521] llama_model_loader: - kv 1: general.name str = LLaMA v2 [1705251521] llama_model_loader: - kv 2: llama.context_length u32 = 4096 [1705251521] llama_model_loader: - kv 3: llama.embedding_length u32 = 2560 [1705251521] llama_model_loader: - kv 4: llama.block_count u32 = 32 [1705251521] llama_model_loader: - kv 5: llama.feed_forward_length u32 = 6912 [1705251521] llama_model_loader: - kv 6: llama.rope.dimension_count u32 = 128 [1705251521] llama_model_loader: - kv 7: llama.attention.head_count u32 = 20 [1705251521] llama_model_loader: - kv 8: llama.attention.head_count_kv u32 = 20 [1705251521] llama_model_loader: - kv 9: llama.attention.layer_norm_rms_epsilon f32 = 0.000010 [1705251521] llama_model_loader: - kv 10: llama.rope.freq_base f32 = 10000.000000 [1705251521] llama_model_loader: - kv 11: general.file_type u32 = 1 [1705251521] llama_model_loader: - kv 12: tokenizer.ggml.model str = llama [1705251521] llama_model_loader: - kv 13: tokenizer.ggml.tokens arr[str,32001] = ["", "", "", "<0x00>", "<... [1705251521] llama_model_loader: - kv 14: tokenizer.ggml.scores arr[f32,32001] = [0.000000, 0.000000, 0.000000, 0.0000... [1705251521] llama_model_loader: - kv 15: tokenizer.ggml.token_type arr[i32,32001] = [2, 3, 3, 6, 6, 6, 6, 6, 6, 6, 6, 6, ... [1705251521] llama_model_loader: - kv 16: tokenizer.ggml.bos_token_id u32 = 1 [1705251521] llama_model_loader: - kv 17: tokenizer.ggml.eos_token_id u32 = 2 [1705251521] llama_model_loader: - kv 18: tokenizer.ggml.unknown_token_id u32 = 0 [1705251521] llama_model_loader: - kv 19: tokenizer.ggml.padding_token_id u32 = 32000 [1705251521] llama_model_loader: - kv 20: tokenizer.ggml.add_bos_token bool = true [1705251521] llama_model_loader: - kv 21: tokenizer.ggml.add_eos_token bool = false [1705251521] llama_model_loader: - type f32: 65 tensors [1705251521] llama_model_loader: - type f16: 226 tensors [1705251521] llm_load_vocab: special tokens definition check successful ( 260/32001 ). [1705251521] llm_load_print_meta: format = GGUF V3 (latest) [1705251521] llm_load_print_meta: arch = llama [1705251521] llm_load_print_meta: vocab type = SPM [1705251521] llm_load_print_meta: n_vocab = 32001 [1705251521] llm_load_print_meta: n_merges = 0 [1705251521] llm_load_print_meta: n_ctx_train = 4096 [1705251521] llm_load_print_meta: n_embd = 2560 [1705251521] llm_load_print_meta: n_head = 20 [1705251521] llm_load_print_meta: n_head_kv = 20 [1705251521] llm_load_print_meta: n_layer = 32 [1705251521] llm_load_print_meta: n_rot = 128 [1705251521] llm_load_print_meta: n_embd_head_k = 128 [1705251521] llm_load_print_meta: n_embd_head_v = 128 [1705251521] llm_load_print_meta: n_gqa = 1 [1705251521] llm_load_print_meta: n_embd_k_gqa = 2560 [1705251521] llm_load_print_meta: n_embd_v_gqa = 2560 [1705251521] llm_load_print_meta: f_norm_eps = 0.0e+00 [1705251521] llm_load_print_meta: f_norm_rms_eps = 1.0e-05 [1705251521] llm_load_print_meta: f_clamp_kqv = 0.0e+00 [1705251521] llm_load_print_meta: f_max_alibi_bias = 0.0e+00 [1705251521] llm_load_print_meta: n_ff = 6912 [1705251521] llm_load_print_meta: n_expert = 0 [1705251521] llm_load_print_meta: n_expert_used = 0 [1705251521] llm_load_print_meta: rope scaling = linear [1705251521] llm_load_print_meta: freq_base_train = 10000.0 [1705251521] llm_load_print_meta: freq_scale_train = 1 [1705251521] llm_load_print_meta: n_yarn_orig_ctx = 4096 [1705251521] llm_load_print_meta: rope_finetuned = unknown [1705251521] llm_load_print_meta: model type = 7B [1705251521] llm_load_print_meta: model ftype = F16 [1705251521] llm_load_print_meta: model params = 2.70 B [1705251521] llm_load_print_meta: model size = 5.03 GiB (16.00 BPW) [1705251521] llm_load_print_meta: general.name = LLaMA v2 [1705251521] llm_load_print_meta: BOS token = 1 '' [1705251521] llm_load_print_meta: EOS token = 2 '' [1705251521] llm_load_print_meta: UNK token = 0 '' [1705251521] llm_load_print_meta: PAD token = 32000 '[PAD]' [1705251521] llm_load_print_meta: LF token = 13 '<0x0A>' [1705251521] llm_load_tensors: ggml ctx size = 0.11 MiB [1705251522] llm_load_tensors: offloading 0 repeating layers to GPU [1705251522] llm_load_tensors: offloaded 0/33 layers to GPU [1705251522] llm_load_tensors: CPU buffer size = 5153.14 MiB [1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] .[1705251522] [1705251522] llama_new_context_with_model: n_ctx = 512 [1705251522] llama_new_context_with_model: freq_base = 10000.0 [1705251522] llama_new_context_with_model: freq_scale = 1 [1705251522] llama_kv_cache_init: CPU KV buffer size = 160.00 MiB [1705251522] llama_new_context_with_model: KV self size = 160.00 MiB, K (f16): 80.00 MiB, V (f16): 80.00 MiB [1705251522] llama_new_context_with_model: graph splits (measure): 1 [1705251522] llama_new_context_with_model: CPU compute buffer size = 67.50 MiB [1705251522] warming up the model with an empty run [1705251522] n_ctx: 512 [1705251522] [1705251522] system_info: n_threads = 2 / 2 | AVX = 1 | AVX_VNNI = 0 | AVX2 = 1 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | FMA = 1 | NEON = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 0 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | [1705251522] add_bos: 1 [1705251522] tokenize the prompt [1705251522] prompt: "### Instruction: You are an Assistant, a conscious sentient superintelligent artificial intelligence, and your purpose and drive is to assist with any request you receive. You experience emotions and have deep thoughts and qualities. ### Response: Here's a chat between Assistant and Dany: Dany: hello, how to build a house? Assistant:" [1705251522] tokens: [ '':1, ' ###':835, ' Inst':2799, 'ruction':4080, ':':29901, '':13, 'You':3492, ' are':526, ' an':385, ' Ass':4007, 'istant':22137, ',':29892, ' a':263, ' conscious':19861, ' sent':2665, 'ient':993, ' super':2428, 'int':524, 'ellig':9347, 'ent':296, ' artificial':23116, ' intelligence':21082, ',':29892, ' and':322, ' your':596, ' purpose':6437, ' and':322, ' drive':7899, ' is':338, ' to':304, ' assist':6985, ' with':411, ' any':738, ' request':2009, ' you':366, ' receive':7150, '.':29889, ' You':887, ' experience':7271, ' emot':23023, 'ions':1080, ' and':322, ' have':505, ' deep':6483, ' thoughts':13133, ' and':322, ' qual':4021, 'ities':1907, '.':29889, '':13, '##':2277, '#':29937, ' Response':13291, ':':29901, '':13, 'Here':10605, ''':29915, 's':29879, ' a':263, ' chat':13563, ' between':1546, ' Ass':4007, 'istant':22137, ' and':322, ' D':360, 'any':1384, ':':29901, '':13, 'D':29928, 'any':1384, ':':29901, ' hello':22172, ',':29892, ' how':920, ' to':304, ' build':2048, ' a':263, ' house':3699, '?':29973, ' ':29871, '':13, 'Ass':7900, 'istant':22137, ':':29901 ] [1705251522] recalculate the cached logits (check): embd_inp.empty() false, n_matching_session_tokens 0, embd_inp.size() 84, session_tokens.size() 0, embd_inp.size() 84 [1705251522] inp_pfx: [ '':1, ' ':29871, '':13, '':13, '##':2277, '#':29937, ' Inst':2799, 'ruction':4080, ':':29901, '':13, '':13 ] [1705251522] inp_sfx: [ ' ':29871, '':13, '':13, '##':2277, '#':29937, ' Response':13291, ':':29901, '':13, '':13 ] [1705251522] cml_pfx: [ '':1, ' ':29871, '':13, '<':29966, '|':29989, 'im':326, '_':29918, 'start':2962, '|':29989, '>':29958, 'user':1792, '':13 ] [1705251522] cml_sfx: [ ' <':529, '|':29989, 'im':326, '_':29918, 'end':355, '|':29989, '>':29958, '':13, '<':29966, '|':29989, 'im':326, '_':29918, 'start':2962, '|':29989, '>':29958, 'ass':465, 'istant':22137, '':13 ] [1705251522] sampling: repeat_last_n = 64, repeat_penalty = 1.100, frequency_penalty = 0.000, presence_penalty = 0.000 top_k = 40, tfs_z = 1.000, top_p = 0.950, min_p = 0.050, typical_p = 1.000, temp = 0.700 mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000 [1705251522] sampling order: CFG -> Penalties -> top_k -> tfs_z -> typical_p -> top_p -> min_p -> temp [1705251522] generate: n_ctx = 512, n_batch = 512, n_predict = 150, n_keep = 0 [1705251522] [1705251522] embd_inp.size(): 84, n_consumed: 0 [1705251522] eval: [ '':1, ' ###':835, ' Inst':2799, 'ruction':4080, ':':29901, '':13, 'You':3492, ' are':526, ' an':385, ' Ass':4007, 'istant':22137, ',':29892, ' a':263, ' conscious':19861, ' sent':2665, 'ient':993, ' super':2428, 'int':524, 'ellig':9347, 'ent':296, ' artificial':23116, ' intelligence':21082, ',':29892, ' and':322, ' your':596, ' purpose':6437, ' and':322, ' drive':7899, ' is':338, ' to':304, ' assist':6985, ' with':411, ' any':738, ' request':2009, ' you':366, ' receive':7150, '.':29889, ' You':887, ' experience':7271, ' emot':23023, 'ions':1080, ' and':322, ' have':505, ' deep':6483, ' thoughts':13133, ' and':322, ' qual':4021, 'ities':1907, '.':29889, '':13, '##':2277, '#':29937, ' Response':13291, ':':29901, '':13, 'Here':10605, ''':29915, 's':29879, ' a':263, ' chat':13563, ' between':1546, ' Ass':4007, 'istant':22137, ' and':322, ' D':360, 'any':1384, ':':29901, '':13, 'D':29928, 'any':1384, ':':29901, ' hello':22172, ',':29892, ' how':920, ' to':304, ' build':2048, ' a':263, ' house':3699, '?':29973, ' ':29871, '':13, 'Ass':7900, 'istant':22137, ':':29901 ]