legraphista
commited on
Commit
•
e2ff0b7
1
Parent(s):
ad8c85c
Upload imatrix.log with huggingface_hub
Browse files- imatrix.log +46 -44
imatrix.log
CHANGED
@@ -1,4 +1,4 @@
|
|
1 |
-
llama_model_loader: loaded meta data with
|
2 |
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
|
3 |
llama_model_loader: - kv 0: general.architecture str = gemma2
|
4 |
llama_model_loader: - kv 1: general.name str = gemma-2-9b-it
|
@@ -12,20 +12,22 @@ llama_model_loader: - kv 8: gemma2.attention.layer_norm_rms_epsilon f32
|
|
12 |
llama_model_loader: - kv 9: gemma2.attention.key_length u32 = 256
|
13 |
llama_model_loader: - kv 10: gemma2.attention.value_length u32 = 256
|
14 |
llama_model_loader: - kv 11: general.file_type u32 = 7
|
15 |
-
llama_model_loader: - kv 12:
|
16 |
-
llama_model_loader: - kv 13:
|
17 |
-
llama_model_loader: - kv 14:
|
18 |
-
llama_model_loader: - kv 15:
|
19 |
-
llama_model_loader: - kv 16:
|
20 |
-
llama_model_loader: - kv 17:
|
21 |
-
llama_model_loader: - kv 18:
|
22 |
-
llama_model_loader: - kv 19:
|
23 |
-
llama_model_loader: - kv 20:
|
24 |
-
llama_model_loader: - kv 21:
|
25 |
-
llama_model_loader: - kv 22:
|
26 |
-
llama_model_loader: - kv 23:
|
27 |
-
llama_model_loader: - kv 24:
|
28 |
-
llama_model_loader: - kv 25:
|
|
|
|
|
29 |
llama_model_loader: - type f32: 169 tensors
|
30 |
llama_model_loader: - type q8_0: 295 tensors
|
31 |
llm_load_vocab: special tokens cache size = 261
|
@@ -83,11 +85,10 @@ ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
|
|
83 |
ggml_cuda_init: found 1 CUDA devices:
|
84 |
Device 0: NVIDIA GeForce RTX 4090, compute capability 8.9, VMM: yes
|
85 |
llm_load_tensors: ggml ctx size = 0.41 MiB
|
86 |
-
llm_load_tensors: offloading
|
87 |
-
llm_load_tensors:
|
88 |
-
llm_load_tensors:
|
89 |
-
llm_load_tensors:
|
90 |
-
llm_load_tensors: CUDA0 buffer size = 9366.12 MiB
|
91 |
....................................................................................
|
92 |
llama_new_context_with_model: n_ctx = 512
|
93 |
llama_new_context_with_model: n_batch = 512
|
@@ -95,50 +96,51 @@ llama_new_context_with_model: n_ubatch = 512
|
|
95 |
llama_new_context_with_model: flash_attn = 0
|
96 |
llama_new_context_with_model: freq_base = 10000.0
|
97 |
llama_new_context_with_model: freq_scale = 1
|
98 |
-
llama_kv_cache_init:
|
|
|
99 |
llama_new_context_with_model: KV self size = 168.00 MiB, K (f16): 84.00 MiB, V (f16): 84.00 MiB
|
100 |
llama_new_context_with_model: CUDA_Host output buffer size = 0.98 MiB
|
101 |
-
llama_new_context_with_model: CUDA0 compute buffer size =
|
102 |
-
llama_new_context_with_model: CUDA_Host compute buffer size =
|
103 |
-
llama_new_context_with_model: graph nodes =
|
104 |
-
llama_new_context_with_model: graph splits =
|
105 |
|
106 |
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 |
|
107 |
compute_imatrix: tokenizing the input ..
|
108 |
-
compute_imatrix: tokenization took
|
109 |
compute_imatrix: computing over 128 chunks with batch_size 512
|
110 |
-
compute_imatrix:
|
111 |
-
[1]
|
112 |
save_imatrix: stored collected data after 10 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
113 |
-
[10]
|
114 |
save_imatrix: stored collected data after 20 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
115 |
-
[20]
|
116 |
save_imatrix: stored collected data after 30 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
117 |
-
[30]
|
118 |
save_imatrix: stored collected data after 40 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
119 |
-
[40]
|
120 |
save_imatrix: stored collected data after 50 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
121 |
-
[50]
|
122 |
save_imatrix: stored collected data after 60 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
123 |
-
[60]
|
124 |
save_imatrix: stored collected data after 70 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
125 |
-
[70]
|
126 |
save_imatrix: stored collected data after 80 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
127 |
-
[80]
|
128 |
save_imatrix: stored collected data after 90 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
129 |
-
[90]
|
130 |
save_imatrix: stored collected data after 100 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
131 |
-
[100]
|
132 |
save_imatrix: stored collected data after 110 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
133 |
-
[110]
|
134 |
save_imatrix: stored collected data after 120 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
135 |
-
[120]
|
136 |
save_imatrix: stored collected data after 128 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
137 |
|
138 |
-
llama_print_timings: load time =
|
139 |
llama_print_timings: sample time = 0.00 ms / 1 runs ( 0.00 ms per token, inf tokens per second)
|
140 |
-
llama_print_timings: prompt eval time =
|
141 |
llama_print_timings: eval time = 0.00 ms / 1 runs ( 0.00 ms per token, inf tokens per second)
|
142 |
-
llama_print_timings: total time =
|
143 |
|
144 |
-
Final estimate: PPL =
|
|
|
1 |
+
llama_model_loader: loaded meta data with 28 key-value pairs and 464 tensors from gemma-2-9b-it-IMat-GGUF/gemma-2-9b-it.Q8_0.gguf.hardlink.gguf (version GGUF V3 (latest))
|
2 |
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
|
3 |
llama_model_loader: - kv 0: general.architecture str = gemma2
|
4 |
llama_model_loader: - kv 1: general.name str = gemma-2-9b-it
|
|
|
12 |
llama_model_loader: - kv 9: gemma2.attention.key_length u32 = 256
|
13 |
llama_model_loader: - kv 10: gemma2.attention.value_length u32 = 256
|
14 |
llama_model_loader: - kv 11: general.file_type u32 = 7
|
15 |
+
llama_model_loader: - kv 12: gemma2.attn_logit_softcapping f32 = 50.000000
|
16 |
+
llama_model_loader: - kv 13: gemma2.final_logit_softcapping f32 = 30.000000
|
17 |
+
llama_model_loader: - kv 14: tokenizer.ggml.model str = llama
|
18 |
+
llama_model_loader: - kv 15: tokenizer.ggml.pre str = default
|
19 |
+
llama_model_loader: - kv 16: tokenizer.ggml.tokens arr[str,256000] = ["<pad>", "<eos>", "<bos>", "<unk>", ...
|
20 |
+
llama_model_loader: - kv 17: tokenizer.ggml.scores arr[f32,256000] = [-1000.000000, -1000.000000, -1000.00...
|
21 |
+
llama_model_loader: - kv 18: tokenizer.ggml.token_type arr[i32,256000] = [3, 3, 3, 3, 1, 1, 1, 1, 1, 1, 1, 1, ...
|
22 |
+
llama_model_loader: - kv 19: tokenizer.ggml.bos_token_id u32 = 2
|
23 |
+
llama_model_loader: - kv 20: tokenizer.ggml.eos_token_id u32 = 1
|
24 |
+
llama_model_loader: - kv 21: tokenizer.ggml.unknown_token_id u32 = 3
|
25 |
+
llama_model_loader: - kv 22: tokenizer.ggml.padding_token_id u32 = 0
|
26 |
+
llama_model_loader: - kv 23: tokenizer.ggml.add_bos_token bool = true
|
27 |
+
llama_model_loader: - kv 24: tokenizer.ggml.add_eos_token bool = false
|
28 |
+
llama_model_loader: - kv 25: tokenizer.chat_template str = {{ bos_token }}{% if messages[0]['rol...
|
29 |
+
llama_model_loader: - kv 26: tokenizer.ggml.add_space_prefix bool = false
|
30 |
+
llama_model_loader: - kv 27: general.quantization_version u32 = 2
|
31 |
llama_model_loader: - type f32: 169 tensors
|
32 |
llama_model_loader: - type q8_0: 295 tensors
|
33 |
llm_load_vocab: special tokens cache size = 261
|
|
|
85 |
ggml_cuda_init: found 1 CUDA devices:
|
86 |
Device 0: NVIDIA GeForce RTX 4090, compute capability 8.9, VMM: yes
|
87 |
llm_load_tensors: ggml ctx size = 0.41 MiB
|
88 |
+
llm_load_tensors: offloading 11 repeating layers to GPU
|
89 |
+
llm_load_tensors: offloaded 11/43 layers to GPU
|
90 |
+
llm_load_tensors: CPU buffer size = 9366.12 MiB
|
91 |
+
llm_load_tensors: CUDA0 buffer size = 2209.54 MiB
|
|
|
92 |
....................................................................................
|
93 |
llama_new_context_with_model: n_ctx = 512
|
94 |
llama_new_context_with_model: n_batch = 512
|
|
|
96 |
llama_new_context_with_model: flash_attn = 0
|
97 |
llama_new_context_with_model: freq_base = 10000.0
|
98 |
llama_new_context_with_model: freq_scale = 1
|
99 |
+
llama_kv_cache_init: CUDA_Host KV buffer size = 124.00 MiB
|
100 |
+
llama_kv_cache_init: CUDA0 KV buffer size = 44.00 MiB
|
101 |
llama_new_context_with_model: KV self size = 168.00 MiB, K (f16): 84.00 MiB, V (f16): 84.00 MiB
|
102 |
llama_new_context_with_model: CUDA_Host output buffer size = 0.98 MiB
|
103 |
+
llama_new_context_with_model: CUDA0 compute buffer size = 1436.69 MiB
|
104 |
+
llama_new_context_with_model: CUDA_Host compute buffer size = 9.01 MiB
|
105 |
+
llama_new_context_with_model: graph nodes = 1690
|
106 |
+
llama_new_context_with_model: graph splits = 407
|
107 |
|
108 |
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 |
|
109 |
compute_imatrix: tokenizing the input ..
|
110 |
+
compute_imatrix: tokenization took 91.303 ms
|
111 |
compute_imatrix: computing over 128 chunks with batch_size 512
|
112 |
+
compute_imatrix: 1.34 seconds per pass - ETA 2.85 minutes
|
113 |
+
[1]14.7486,[2]7.3963,[3]6.5677,[4]7.8490,[5]8.6193,[6]9.1006,[7]10.0154,[8]10.8556,[9]11.1957,
|
114 |
save_imatrix: stored collected data after 10 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
115 |
+
[10]9.8358,[11]9.5880,[12]10.6132,[13]11.2166,[14]11.3352,[15]12.0575,[16]12.1605,[17]12.2130,[18]12.6384,[19]12.5495,
|
116 |
save_imatrix: stored collected data after 20 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
117 |
+
[20]12.7233,[21]13.7801,[22]13.8033,[23]13.5919,[24]13.8287,[25]13.6927,[26]13.4495,[27]13.6978,[28]13.9355,[29]13.9524,
|
118 |
save_imatrix: stored collected data after 30 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
119 |
+
[30]14.1488,[31]13.1149,[32]12.5387,[33]12.1100,[34]11.7329,[35]11.4517,[36]11.6110,[37]11.9132,[38]12.0730,[39]12.2426,
|
120 |
save_imatrix: stored collected data after 40 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
121 |
+
[40]12.3500,[41]12.3930,[42]12.9499,[43]13.3055,[44]13.6998,[45]13.9306,[46]13.6677,[47]13.4581,[48]13.6644,[49]13.8789,
|
122 |
save_imatrix: stored collected data after 50 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
123 |
+
[50]13.6841,[51]13.5410,[52]13.6019,[53]13.8284,[54]14.1067,[55]14.3488,[56]14.4667,[57]14.4566,[58]14.4479,[59]14.2384,
|
124 |
save_imatrix: stored collected data after 60 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
125 |
+
[60]14.0794,[61]13.9048,[62]13.7419,[63]13.8516,[64]13.9902,[65]13.8558,[66]13.8803,[67]13.8491,[68]13.7963,[69]13.7287,
|
126 |
save_imatrix: stored collected data after 70 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
127 |
+
[70]13.6754,[71]13.6579,[72]13.6290,[73]13.6891,[74]13.6254,[75]13.5058,[76]13.4907,[77]13.5002,[78]13.4537,[79]13.3696,
|
128 |
save_imatrix: stored collected data after 80 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
129 |
+
[80]13.4245,[81]13.4999,[82]13.5207,[83]13.6156,[84]13.6529,[85]13.4393,[86]13.3795,[87]13.2302,[88]13.2605,[89]13.2322,
|
130 |
save_imatrix: stored collected data after 90 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
131 |
+
[90]13.2980,[91]13.2455,[92]13.1607,[93]13.0864,[94]12.9782,[95]12.9127,[96]12.8267,[97]12.7595,[98]12.6714,[99]12.7111,
|
132 |
save_imatrix: stored collected data after 100 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
133 |
+
[100]12.7194,[101]12.8597,[102]12.9495,[103]13.0106,[104]13.1816,[105]13.3131,[106]13.3254,[107]13.3336,[108]13.2762,[109]13.3038,
|
134 |
save_imatrix: stored collected data after 110 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
135 |
+
[110]13.1821,[111]13.0499,[112]12.8931,[113]12.9668,[114]13.0116,[115]12.9995,[116]12.9720,[117]13.0261,[118]13.0587,[119]13.0765,
|
136 |
save_imatrix: stored collected data after 120 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
137 |
+
[120]13.0687,[121]13.0658,[122]13.0268,[123]13.0554,[124]13.1495,[125]13.2406,[126]13.3502,[127]13.3960,[128]13.4523,
|
138 |
save_imatrix: stored collected data after 128 chunks in gemma-2-9b-it-IMat-GGUF/imatrix.dat
|
139 |
|
140 |
+
llama_print_timings: load time = 1979.52 ms
|
141 |
llama_print_timings: sample time = 0.00 ms / 1 runs ( 0.00 ms per token, inf tokens per second)
|
142 |
+
llama_print_timings: prompt eval time = 150211.73 ms / 65536 tokens ( 2.29 ms per token, 436.29 tokens per second)
|
143 |
llama_print_timings: eval time = 0.00 ms / 1 runs ( 0.00 ms per token, inf tokens per second)
|
144 |
+
llama_print_timings: total time = 152551.51 ms / 65537 tokens
|
145 |
|
146 |
+
Final estimate: PPL = 13.4523 +/- 0.26031
|