Guest User

Untitled

a guest
Mar 22nd, 2026
369
0
Never
2
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 7.45 KB | None | 0 0
  1. > ./infer --prompt "Explain quantum computing" --tokens 100 Sun Mar 22 14:11:31 2026
  2. [metal] Device: Apple M5 Pro
  3. [metal] Shader compile: 2 ms
  4. [metal] GPU attention buffers: 15 KV caches (16.8 MB each), scores buf 134.2 MB
  5. [metal] Delta-net GPU buffers: 45 layers (195.4 MB state + 0.2 MB scratch)
  6. [metal] Inference pipelines ready (multi-expert[8] + shared buffers allocated)
  7. === Qwen3.5-397B-A17B Metal Inference Engine ===
  8. Model: /Users/redacted/.cache/huggingface/hub/models--mlx-community--Qwen3.5-397B-A17B-4bit/snapshots/39159bd8aa74f5c8446d2b2dc584f62bb51cb0d3
  9. Weights: model_weights.bin
  10. Manifest: model_weights.json
  11. Vocab: vocab.bin
  12. K: 4 experts/layer
  13. Quant: 4-bit experts (7077888 bytes each)
  14. Linear: fused GPU delta-net
  15. Tokens: 100
  16. Cache: 0 entries (disabled)
  17. [manifest] Loaded 2092 tensors from model_weights.json
  18. [weights] mmap'd 5.52 GB from model_weights.bin
  19. [metal] Weight file wrapped as Metal buffer (5.52 GB)
  20. [vocab] Loaded 248077 tokens
  21. bpe_load: 248044 vocab, 247587 merges, 33 added tokens
  22. Tokens (4): [814, 20139, 29144, 23470]
  23. [prompt] 4 tokens: 814 20139 29144 23470
  24. [experts] 60/60 packed layer files available (mmap'd)
  25. [tiered-io] Cold fds (F_NOCACHE) + warm fds (page cached) active
  26. [warmup] Page cache hint: 5.2 ms
  27. [init] Setup: 69.8 ms
  28.  
  29. --- Generating 100 tokens ---
  30. [prefill] batch embed 4 tokens: 1.2 ms
  31. [cache] Pre-computed weight pointers for 60 layers
  32. [prefill] 3/4 tokens: 1707 ms (first: 1220 ms, rest avg: 244 ms)
  33. [debug] Top 5 logits (next_token=303):
  34. token 1040 ("Ġlike") logit=17.6052
  35. token 303 ("Ġin") logit=20.5380
  36. token 248046 ("<|im_end|>") logit=17.1903
  37. token 310 ("Ġto") logit=19.7185
  38. token 430 ("Ġas") logit=17.5620
  39. [debug] hidden rms after final_norm=1.7007, logits rms=2.2827
  40. [ttft] 2010 ms (prefill 4 tokens + lm_head 3 ms)
  41.  
  42. --- Output ---
  43. ĠinĠsimple [gen 1/100] token_id=4145 (214 ms, 4.67 tok/s)
  44. Ġterms [gen 2/100] token_id=3665 (184 ms, 5.44 tok/s)
  45. . [gen 3/100] token_id=13 (184 ms, 5.42 tok/s)
  46. ĊĊ [gen 4/100] token_id=271 (180 ms, 5.55 tok/s)
  47. <think> [gen 5/100] token_id=248068 (173 ms, 5.77 tok/s)
  48. ĊĊ [gen 6/100] token_id=271 (164 ms, 6.11 tok/s)
  49. </think> [gen 7/100] token_id=248069 (150 ms, 6.66 tok/s)
  50. ĊĊ [gen 8/100] token_id=271 (152 ms, 6.59 tok/s)
  51. Here [gen 9/100] token_id=8160 (141 ms, 7.10 tok/s)
  52. Ġis [gen 10/100] token_id=369 (168 ms, 5.95 tok/s)
  53. Ġan [gen 11/100] token_id=449 (186 ms, 5.39 tok/s)
  54. Ġexplanation [gen 12/100] token_id=15673 (185 ms, 5.40 tok/s)
  55. Ġof [gen 13/100] token_id=314 (182 ms, 5.50 tok/s)
  56. Ġquantum [gen 14/100] token_id=29144 (159 ms, 6.31 tok/s)
  57. Ġcomputing [gen 15/100] token_id=23470 (180 ms, 5.56 tok/s)
  58. Ġin [gen 16/100] token_id=303 (156 ms, 6.43 tok/s)
  59. Ġsimple [gen 17/100] token_id=4145 (142 ms, 7.06 tok/s)
  60. Ġterms [gen 18/100] token_id=3665 (137 ms, 7.32 tok/s)
  61. , [gen 19/100] token_id=11 (135 ms, 7.42 tok/s)
  62. Ġbreaking [gen 20/100] token_id=14293 (182 ms, 5.50 tok/s)
  63. Ġdown [gen 21/100] token_id=1441 (184 ms, 5.42 tok/s)
  64. Ġthe [gen 22/100] token_id=279 (147 ms, 6.80 tok/s)
  65. Ġcomplex [gen 23/100] token_id=6150 (180 ms, 5.55 tok/s)
  66. Ġj [gen 24/100] token_id=492 (165 ms, 6.05 tok/s)
  67. argon [gen 25/100] token_id=68421 (189 ms, 5.29 tok/s)
  68. Ġinto [gen 26/100] token_id=1083 (135 ms, 7.39 tok/s)
  69. Ġeasy [gen 27/100] token_id=3999 (150 ms, 6.65 tok/s)
  70. -to [gen 28/100] token_id=4534 (167 ms, 6.00 tok/s)
  71. -under [gen 29/100] token_id=68695 (178 ms, 5.63 tok/s)
  72. stand [gen 30/100] token_id=2598 (145 ms, 6.88 tok/s)
  73. Ġconcepts [gen 31/100] token_id=18369 (136 ms, 7.34 tok/s)
  74. . [gen 32/100] token_id=13 (154 ms, 6.50 tok/s)
  75. ĊĊ [gen 33/100] token_id=271 (137 ms, 7.27 tok/s)
  76. ### [gen 34/100] token_id=13962 (127 ms, 7.88 tok/s)
  77. Ġ [gen 35/100] token_id=220 (127 ms, 7.90 tok/s)
  78. 1 [gen 36/100] token_id=16 (152 ms, 6.58 tok/s)
  79. . [gen 37/100] token_id=13 (151 ms, 6.61 tok/s)
  80. ĠThe [gen 38/100] token_id=561 (124 ms, 8.07 tok/s)
  81. ĠBasic [gen 39/100] token_id=14204 (134 ms, 7.48 tok/s)
  82. ĠDifference [gen 40/100] token_id=53230 (129 ms, 7.76 tok/s)
  83. : [gen 41/100] token_id=25 (148 ms, 6.75 tok/s)
  84. ĠSwitch [gen 42/100] token_id=15129 (136 ms, 7.37 tok/s)
  85. es [gen 43/100] token_id=287 (163 ms, 6.15 tok/s)
  86. Ġvs [gen 44/100] token_id=5972 (133 ms, 7.51 tok/s)
  87. . [gen 45/100] token_id=13 (162 ms, 6.17 tok/s)
  88. ĠSp [gen 46/100] token_id=2989 (146 ms, 6.86 tok/s)
  89. inning [gen 47/100] token_id=19167 (166 ms, 6.01 tok/s)
  90. ĠCoins [gen 48/100] token_id=59692 (130 ms, 7.69 tok/s)
  91. Ċ [gen 49/100] token_id=198 (164 ms, 6.11 tok/s)
  92. To [gen 50/100] token_id=1206 (132 ms, 7.56 tok/s)
  93. Ġunderstand [gen 51/100] token_id=3418 (154 ms, 6.50 tok/s)
  94. Ġquantum [gen 52/100] token_id=29144 (150 ms, 6.68 tok/s)
  95. Ġcomputing [gen 53/100] token_id=23470 (173 ms, 5.78 tok/s)
  96. , [gen 54/100] token_id=11 (167 ms, 5.98 tok/s)
  97. Ġfirst [gen 55/100] token_id=1118 (172 ms, 5.82 tok/s)
  98. Ġthink [gen 56/100] token_id=1683 (166 ms, 6.01 tok/s)
  99. Ġabout [gen 57/100] token_id=883 (164 ms, 6.09 tok/s)
  100. Ġhow [gen 58/100] token_id=1204 (138 ms, 7.27 tok/s)
  101. Ġyour [gen 59/100] token_id=678 (157 ms, 6.36 tok/s)
  102. Ġcurrent [gen 60/100] token_id=1428 (170 ms, 5.88 tok/s)
  103. Ġcomputer [gen 61/100] token_id=6165 (146 ms, 6.87 tok/s)
  104. Ġ( [gen 62/100] token_id=318 (161 ms, 6.19 tok/s)
  105. a [gen 63/100] token_id=64 (168 ms, 5.95 tok/s)
  106. Ġclassical [gen 64/100] token_id=27882 (148 ms, 6.74 tok/s)
  107. Ġcomputer [gen 65/100] token_id=6165 (144 ms, 6.93 tok/s)
  108. ) [gen 66/100] token_id=8 (154 ms, 6.50 tok/s)
  109. Ġworks [gen 67/100] token_id=4138 (152 ms, 6.57 tok/s)
  110. . [gen 68/100] token_id=13 (151 ms, 6.63 tok/s)
  111. Ċ [gen 69/100] token_id=198 (159 ms, 6.30 tok/s)
  112. * [gen 70/100] token_id=9 (188 ms, 5.33 tok/s)
  113. ĠĠ [gen 71/100] token_id=256 (160 ms, 6.24 tok/s)
  114. Ġ** [gen 72/100] token_id=2972 (128 ms, 7.79 tok/s)
  115. Class [gen 73/100] token_id=1890 (127 ms, 7.85 tok/s)
  116. ical [gen 74/100] token_id=908 (131 ms, 7.63 tok/s)
  117. ĠComputers [gen 75/100] token_id=66133 (130 ms, 7.72 tok/s)
  118. :** [gen 76/100] token_id=64700 (123 ms, 8.10 tok/s)
  119. ĠImagine [gen 77/100] token_id=36541 (144 ms, 6.97 tok/s)
  120. Ġa [gen 78/100] token_id=264 (123 ms, 8.14 tok/s)
  121. Ġlight [gen 79/100] token_id=3000 (136 ms, 7.34 tok/s)
  122. Ġswitch [gen 80/100] token_id=3289 (133 ms, 7.49 tok/s)
  123. . [gen 81/100] token_id=13 (151 ms, 6.62 tok/s)
  124. ĠIt [gen 82/100] token_id=1049 (118 ms, 8.45 tok/s)
  125. Ġcan [gen 83/100] token_id=628 (151 ms, 6.61 tok/s)
  126. Ġbe [gen 84/100] token_id=381 (158 ms, 6.33 tok/s)
  127. Ġeither [gen 85/100] token_id=2892 (149 ms, 6.71 tok/s)
  128. Ġ** [gen 86/100] token_id=2972 (153 ms, 6.54 tok/s)
  129. ON [gen 87/100] token_id=691 (159 ms, 6.28 tok/s)
  130. ** [gen 88/100] token_id=332 (158 ms, 6.34 tok/s)
  131. Ġ( [gen 89/100] token_id=318 (152 ms, 6.59 tok/s)
  132. 1 [gen 90/100] token_id=16 (134 ms, 7.49 tok/s)
  133. ) [gen 91/100] token_id=8 (156 ms, 6.39 tok/s)
  134. Ġor [gen 92/100] token_id=466 (129 ms, 7.77 tok/s)
  135. Ġ** [gen 93/100] token_id=2972 (164 ms, 6.12 tok/s)
  136. OFF [gen 94/100] token_id=26200 (143 ms, 7.00 tok/s)
  137. ** [gen 95/100] token_id=332 (135 ms, 7.41 tok/s)
  138. Ġ( [gen 96/100] token_id=318 (143 ms, 6.97 tok/s)
  139. 0 [gen 97/100] token_id=15 (148 ms, 6.77 tok/s)
  140. ). [gen 98/100] token_id=553 (138 ms, 7.24 tok/s)
  141. ĠEvery [gen 99/100] token_id=6983 (123 ms, 8.14 tok/s)
  142.  
  143.  
  144. --- Statistics ---
  145. Total time: 17.1 s
  146. TTFT: 2010 ms
  147. Tokens: 100 generated
  148. Generation: 15.1 s (6.55 tok/s)
  149. Config: K=4 experts, 60 layers
Advertisement
Comments
  • User was banned
  • User was banned
Add Comment
Please, Sign In to add comment