API tools faq
paste
Guest User

Untitled

a guest
Aug 4th, 2025
19
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 16.89 KB | None | 0 0
raw download clone embed print report
  1. Log start
  2. llama_model_loader: loaded meta data with 45 key-value pairs and 579 tensors from Qwen3-30B-A3B-Thinking-2507-Q4_K_S.gguf (version GGUF V3 (latest))
  3. llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
  4. llama_model_loader: - kv 0: general.architecture str = qwen3moe
  5. llama_model_loader: - kv 1: general.type str = model
  6. llama_model_loader: - kv 2: general.name str = Qwen3-30B-A3B-Thinking-2507
  7. llama_model_loader: - kv 3: general.version str = 2507
  8. llama_model_loader: - kv 4: general.finetune str = Thinking
  9. llama_model_loader: - kv 5: general.basename str = Qwen3-30B-A3B-Thinking-2507
  10. llama_model_loader: - kv 6: general.quantized_by str = Unsloth
  11. llama_model_loader: - kv 7: general.size_label str = 30B-A3B
  12. llama_model_loader: - kv 8: general.license str = apache-2.0
  13. llama_model_loader: - kv 9: general.license.link str = https://huggingface.co/Qwen/Qwen3-30B...
  14. llama_model_loader: - kv 10: general.repo_url str = https://huggingface.co/unsloth
  15. llama_model_loader: - kv 11: general.base_model.count u32 = 1
  16. llama_model_loader: - kv 12: general.base_model.0.name str = Qwen3 30B A3B Thinking 2507
  17. llama_model_loader: - kv 13: general.base_model.0.version str = 2507
  18. llama_model_loader: - kv 14: general.base_model.0.organization str = Qwen
  19. llama_model_loader: - kv 15: general.base_model.0.repo_url str = https://huggingface.co/Qwen/Qwen3-30B...
  20. llama_model_loader: - kv 16: general.tags arr[str,2] = ["unsloth", "text-generation"]
  21. llama_model_loader: - kv 17: qwen3moe.block_count u32 = 48
  22. llama_model_loader: - kv 18: qwen3moe.context_length u32 = 262144
  23. llama_model_loader: - kv 19: qwen3moe.embedding_length u32 = 2048
  24. llama_model_loader: - kv 20: qwen3moe.feed_forward_length u32 = 6144
  25. llama_model_loader: - kv 21: qwen3moe.attention.head_count u32 = 32
  26. llama_model_loader: - kv 22: qwen3moe.attention.head_count_kv u32 = 4
  27. llama_model_loader: - kv 23: qwen3moe.rope.freq_base f32 = 10000000.000000
  28. llama_model_loader: - kv 24: qwen3moe.attention.layer_norm_rms_epsilon f32 = 0.000001
  29. llama_model_loader: - kv 25: qwen3moe.expert_used_count u32 = 8
  30. llama_model_loader: - kv 26: qwen3moe.attention.key_length u32 = 128
  31. llama_model_loader: - kv 27: qwen3moe.attention.value_length u32 = 128
  32. llama_model_loader: - kv 28: qwen3moe.expert_count u32 = 128
  33. llama_model_loader: - kv 29: qwen3moe.expert_feed_forward_length u32 = 768
  34. llama_model_loader: - kv 30: tokenizer.ggml.model str = gpt2
  35. llama_model_loader: - kv 31: tokenizer.ggml.pre str = qwen2
  36. llama_model_loader: - kv 32: tokenizer.ggml.tokens arr[str,151936] = ["!", "\"", "#", "$", "%", "&", "'", ...
  37. llama_model_loader: - kv 33: tokenizer.ggml.token_type arr[i32,151936] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
  38. llama_model_loader: - kv 34: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
  39. llama_model_loader: - kv 35: tokenizer.ggml.eos_token_id u32 = 151645
  40. llama_model_loader: - kv 36: tokenizer.ggml.padding_token_id u32 = 151654
  41. llama_model_loader: - kv 37: tokenizer.ggml.add_bos_token bool = false
  42. llama_model_loader: - kv 38: tokenizer.chat_template str = {%- if tools %}\n {{- '<|im_start|>...
  43. llama_model_loader: - kv 39: general.quantization_version u32 = 2
  44. llama_model_loader: - kv 40: general.file_type u32 = 14
  45. llama_model_loader: - kv 41: quantize.imatrix.file str = Qwen3-30B-A3B-Thinking-2507-GGUF/imat...
  46. llama_model_loader: - kv 42: quantize.imatrix.dataset str = unsloth_calibration_Qwen3-30B-A3B-Thi...
  47. llama_model_loader: - kv 43: quantize.imatrix.entries_count u32 = 384
  48. llama_model_loader: - kv 44: quantize.imatrix.chunks_count u32 = 684
  49. llama_model_loader: - type f32: 241 tensors
  50. llama_model_loader: - type q4_K: 327 tensors
  51. llama_model_loader: - type q5_K: 10 tensors
  52. llama_model_loader: - type q6_K: 1 tensors
  53. llm_load_vocab: special tokens cache size = 26
  54. llm_load_vocab: token to piece cache size = 0.9311 MB
  55. llm_load_print_meta: format = GGUF V3 (latest)
  56. llm_load_print_meta: arch = qwen3moe
  57. llm_load_print_meta: vocab type = BPE
  58. llm_load_print_meta: n_vocab = 151936
  59. llm_load_print_meta: n_merges = 151387
  60. llm_load_print_meta: vocab_only = 0
  61. llm_load_print_meta: n_ctx_train = 262144
  62. llm_load_print_meta: n_embd = 2048
  63. llm_load_print_meta: n_layer = 48
  64. llm_load_print_meta: n_head = 32
  65. llm_load_print_meta: n_head_kv = 4
  66. llm_load_print_meta: n_rot = 128
  67. llm_load_print_meta: n_swa = 0
  68. llm_load_print_meta: n_swa_pattern = 1
  69. llm_load_print_meta: n_embd_head_k = 128
  70. llm_load_print_meta: n_embd_head_v = 128
  71. llm_load_print_meta: n_gqa = 8
  72. llm_load_print_meta: n_embd_k_gqa = 512
  73. llm_load_print_meta: n_embd_v_gqa = 512
  74. llm_load_print_meta: f_norm_eps = 0.0e+00
  75. llm_load_print_meta: f_norm_rms_eps = 1.0e-06
  76. llm_load_print_meta: f_clamp_kqv = 0.0e+00
  77. llm_load_print_meta: f_max_alibi_bias = 0.0e+00
  78. llm_load_print_meta: f_logit_scale = 0.0e+00
  79. llm_load_print_meta: n_ff = 6144
  80. llm_load_print_meta: n_expert = 128
  81. llm_load_print_meta: n_expert_used = 8
  82. llm_load_print_meta: causal attn = 1
  83. llm_load_print_meta: pooling type = 0
  84. llm_load_print_meta: rope type = 2
  85. llm_load_print_meta: rope scaling = linear
  86. llm_load_print_meta: freq_base_train = 10000000.0
  87. llm_load_print_meta: freq_scale_train = 1
  88. llm_load_print_meta: n_ctx_orig_yarn = 262144
  89. llm_load_print_meta: rope_finetuned = unknown
  90. llm_load_print_meta: ssm_d_conv = 0
  91. llm_load_print_meta: ssm_d_inner = 0
  92. llm_load_print_meta: ssm_d_state = 0
  93. llm_load_print_meta: ssm_dt_rank = 0
  94. llm_load_print_meta: model type = ?B
  95. llm_load_print_meta: model ftype = Q4_K - Small
  96. llm_load_print_meta: model params = 30.532 B
  97. llm_load_print_meta: model size = 16.252 GiB (4.572 BPW)
  98. llm_load_print_meta: repeating layers = 15.851 GiB (4.552 BPW, 29.910 B parameters)
  99. llm_load_print_meta: general.name = Qwen3-30B-A3B-Thinking-2507
  100. llm_load_print_meta: BOS token = 11 ','
  101. llm_load_print_meta: EOS token = 151645 '<|im_end|>'
  102. llm_load_print_meta: PAD token = 151654 '<|vision_pad|>'
  103. llm_load_print_meta: LF token = 148848 'ÄĬ'
  104. llm_load_print_meta: EOT token = 151645 '<|im_end|>'
  105. llm_load_print_meta: max token length = 256
  106. llm_load_print_meta: n_ff_exp = 768
  107. llm_load_tensors: ggml ctx size = 0.25 MiB
  108. llm_load_tensors: CPU buffer size = 16641.65 MiB
  109. ....................................................................................................
  110. llama_new_context_with_model: n_ctx = 32768
  111. llama_new_context_with_model: n_batch = 2048
  112. llama_new_context_with_model: n_ubatch = 512
  113. llama_new_context_with_model: flash_attn = 0
  114. llama_new_context_with_model: mla_attn = 0
  115. llama_new_context_with_model: attn_max_b = 0
  116. llama_new_context_with_model: fused_moe = 0
  117. llama_new_context_with_model: ser = -1, 0
  118. llama_new_context_with_model: freq_base = 10000000.0
  119. llama_new_context_with_model: freq_scale = 1
  120. llama_kv_cache_init: CPU KV buffer size = 3072.00 MiB
  121. llama_new_context_with_model: KV self size = 3072.00 MiB, K (f16): 1536.00 MiB, V (f16): 1536.00 MiB
  122. llama_new_context_with_model: CPU output buffer size = 0.58 MiB
  123. llama_new_context_with_model: CPU compute buffer size = 2136.01 MiB
  124. llama_new_context_with_model: graph nodes = 2165
  125. llama_new_context_with_model: graph splits = 578
  126. llama_model_loader: loaded meta data with 32 key-value pairs and 310 tensors from Qwen3-0.6B-UD-Q5_K_XL.gguf (version GGUF V3 (latest))
  127. llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
  128. llama_model_loader: - kv 0: general.architecture str = qwen3
  129. llama_model_loader: - kv 1: general.type str = model
  130. llama_model_loader: - kv 2: general.name str = Qwen3-0.6B
  131. llama_model_loader: - kv 3: general.basename str = Qwen3-0.6B
  132. llama_model_loader: - kv 4: general.quantized_by str = Unsloth
  133. llama_model_loader: - kv 5: general.size_label str = 0.6B
  134. llama_model_loader: - kv 6: general.repo_url str = https://huggingface.co/unsloth
  135. llama_model_loader: - kv 7: qwen3.block_count u32 = 28
  136. llama_model_loader: - kv 8: qwen3.context_length u32 = 40960
  137. llama_model_loader: - kv 9: qwen3.embedding_length u32 = 1024
  138. llama_model_loader: - kv 10: qwen3.feed_forward_length u32 = 3072
  139. llama_model_loader: - kv 11: qwen3.attention.head_count u32 = 16
  140. llama_model_loader: - kv 12: qwen3.attention.head_count_kv u32 = 8
  141. llama_model_loader: - kv 13: qwen3.rope.freq_base f32 = 1000000.000000
  142. llama_model_loader: - kv 14: qwen3.attention.layer_norm_rms_epsilon f32 = 0.000001
  143. llama_model_loader: - kv 15: qwen3.attention.key_length u32 = 128
  144. llama_model_loader: - kv 16: qwen3.attention.value_length u32 = 128
  145. llama_model_loader: - kv 17: tokenizer.ggml.model str = gpt2
  146. llama_model_loader: - kv 18: tokenizer.ggml.pre str = qwen2
  147. llama_model_loader: - kv 19: tokenizer.ggml.tokens arr[str,151936] = ["!", "\"", "#", "$", "%", "&", "'", ...
  148. llama_model_loader: - kv 20: tokenizer.ggml.token_type arr[i32,151936] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
  149. llama_model_loader: - kv 21: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
  150. llama_model_loader: - kv 22: tokenizer.ggml.eos_token_id u32 = 151645
  151. llama_model_loader: - kv 23: tokenizer.ggml.padding_token_id u32 = 151654
  152. llama_model_loader: - kv 24: tokenizer.ggml.add_bos_token bool = false
  153. llama_model_loader: - kv 25: tokenizer.chat_template str = {%- if tools %}\n {{- '<|im_start|>...
  154. llama_model_loader: - kv 26: general.quantization_version u32 = 2
  155. llama_model_loader: - kv 27: general.file_type u32 = 17
  156. llama_model_loader: - kv 28: quantize.imatrix.file str = Qwen3-0.6B-GGUF/imatrix_unsloth.dat
  157. llama_model_loader: - kv 29: quantize.imatrix.dataset str = unsloth_calibration_Qwen3-0.6B.txt
  158. llama_model_loader: - kv 30: quantize.imatrix.entries_count u32 = 196
  159. llama_model_loader: - kv 31: quantize.imatrix.chunks_count u32 = 688
  160. llama_model_loader: - type f32: 113 tensors
  161. llama_model_loader: - type q8_0: 1 tensors
  162. llama_model_loader: - type q4_K: 20 tensors
  163. llama_model_loader: - type q5_K: 120 tensors
  164. llama_model_loader: - type q6_K: 56 tensors
  165. llm_load_vocab: special tokens cache size = 26
  166. llm_load_vocab: token to piece cache size = 0.9311 MB
  167. llm_load_print_meta: format = GGUF V3 (latest)
  168. llm_load_print_meta: arch = qwen3
  169. llm_load_print_meta: vocab type = BPE
  170. llm_load_print_meta: n_vocab = 151936
  171. llm_load_print_meta: n_merges = 151387
  172. llm_load_print_meta: vocab_only = 0
  173. llm_load_print_meta: n_ctx_train = 40960
  174. llm_load_print_meta: n_embd = 1024
  175. llm_load_print_meta: n_layer = 28
  176. llm_load_print_meta: n_head = 16
  177. llm_load_print_meta: n_head_kv = 8
  178. llm_load_print_meta: n_rot = 128
  179. llm_load_print_meta: n_swa = 0
  180. llm_load_print_meta: n_swa_pattern = 1
  181. llm_load_print_meta: n_embd_head_k = 128
  182. llm_load_print_meta: n_embd_head_v = 128
  183. llm_load_print_meta: n_gqa = 2
  184. llm_load_print_meta: n_embd_k_gqa = 1024
  185. llm_load_print_meta: n_embd_v_gqa = 1024
  186. llm_load_print_meta: f_norm_eps = 0.0e+00
  187. llm_load_print_meta: f_norm_rms_eps = 1.0e-06
  188. llm_load_print_meta: f_clamp_kqv = 0.0e+00
  189. llm_load_print_meta: f_max_alibi_bias = 0.0e+00
  190. llm_load_print_meta: f_logit_scale = 0.0e+00
  191. llm_load_print_meta: n_ff = 3072
  192. llm_load_print_meta: n_expert = 0
  193. llm_load_print_meta: n_expert_used = 0
  194. llm_load_print_meta: causal attn = 1
  195. llm_load_print_meta: pooling type = 0
  196. llm_load_print_meta: rope type = 2
  197. llm_load_print_meta: rope scaling = linear
  198. llm_load_print_meta: freq_base_train = 1000000.0
  199. llm_load_print_meta: freq_scale_train = 1
  200. llm_load_print_meta: n_ctx_orig_yarn = 40960
  201. llm_load_print_meta: rope_finetuned = unknown
  202. llm_load_print_meta: ssm_d_conv = 0
  203. llm_load_print_meta: ssm_d_inner = 0
  204. llm_load_print_meta: ssm_d_state = 0
  205. llm_load_print_meta: ssm_dt_rank = 0
  206. llm_load_print_meta: model type = ?B
  207. llm_load_print_meta: model ftype = Q5_K - Medium
  208. llm_load_print_meta: model params = 596.050 M
  209. llm_load_print_meta: model size = 420.026 MiB (5.911 BPW)
  210. llm_load_print_meta: general.name = Qwen3-0.6B
  211. llm_load_print_meta: BOS token = 11 ','
  212. llm_load_print_meta: EOS token = 151645 '<|im_end|>'
  213. llm_load_print_meta: PAD token = 151654 '<|vision_pad|>'
  214. llm_load_print_meta: LF token = 148848 'ÄĬ'
  215. llm_load_print_meta: EOT token = 151645 '<|im_end|>'
  216. llm_load_print_meta: max token length = 256
  217. llm_load_tensors: ggml ctx size = 0.14 MiB
  218. llm_load_tensors: CPU buffer size = 420.03 MiB
  219. ..........................................................
  220. llama_new_context_with_model: n_ctx = 32768
  221. llama_new_context_with_model: n_batch = 2048
  222. llama_new_context_with_model: n_ubatch = 512
  223. llama_new_context_with_model: flash_attn = 0
  224. llama_new_context_with_model: mla_attn = 0
  225. llama_new_context_with_model: attn_max_b = 0
  226. llama_new_context_with_model: fused_moe = 0
  227. llama_new_context_with_model: ser = -1, 0
  228. llama_new_context_with_model: freq_base = 1000000.0
  229. llama_new_context_with_model: freq_scale = 1
  230. llama_kv_cache_init: CPU KV buffer size = 3584.00 MiB
  231. llama_new_context_with_model: KV self size = 3584.00 MiB, K (f16): 1792.00 MiB, V (f16): 1792.00 MiB
  232. llama_new_context_with_model: CPU output buffer size = 0.58 MiB
  233. llama_new_context_with_model: CPU compute buffer size = 1100.01 MiB
  234. llama_new_context_with_model: graph nodes = 873
  235. llama_new_context_with_model: graph splits = 394
  236.  
  237.  
  238. /build/source/src/llama.cpp:18273: GGML_ASSERT(n_tokens_all <= cparams.n_batch) failed
  239. /nix/store/ila9g3xmkicpfgpvyx9db2cwv23ng9ni-llama-cpp-blas-0.0.0/lib/libggml.so(+0x1e4eb)[0x7f58f561e4eb]
  240. /nix/store/ila9g3xmkicpfgpvyx9db2cwv23ng9ni-llama-cpp-blas-0.0.0/lib/libggml.so(ggml_abort+0x15f)[0x7f58f562010f]
  241. /nix/store/ila9g3xmkicpfgpvyx9db2cwv23ng9ni-llama-cpp-blas-0.0.0/lib/libllama.so(llama_decode+0x1976)[0x7f58f6185ab6]
  242. llama-speculative[0x41cc7c]
  243. /nix/store/0wydilnf1c9vznywsvxqnaing4wraaxp-glibc-2.39-52/lib/libc.so.6(+0x2a14e)[0x7f58f503314e]
  244. /nix/store/0wydilnf1c9vznywsvxqnaing4wraaxp-glibc-2.39-52/lib/libc.so.6(__libc_start_main+0x89)[0x7f58f5033209]
  245. llama-speculative[0x421515]
  246. Aborted (core dumped) llama-speculative -m Qwen3-30B-A3B-Thinking-2507-Q4_K_S.gguf -md Qwen3-0.6B-UD-Q5_K_XL.gguf -c 32768
Advertisement
Add Comment
Please, Sign In to add comment
Public Pastes
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!