guell00 commited on
Commit
911ef88
·
verified ·
1 Parent(s): bd6894d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +252 -698
README.md CHANGED
@@ -1,7 +1,4 @@
1
  ---
2
- base_model: Jackrong/Qwopus3.5-9B-Coder
3
- datasets:
4
- - guell00/Coder-max
5
  language:
6
  - pt
7
  - en
@@ -17,156 +14,16 @@ pipeline_tag: text-generation
17
  tags:
18
  - code
19
  - coding
20
- - coder
21
- - qwen
22
- - qwen3.5
23
  - gguf
24
  - unsloth
25
  - lora
26
- - sft
27
  - html
28
- - css
29
  - javascript
30
  - threejs
31
- - canvas
32
  - python
33
  - multilingual
34
  ---
35
 
36
- <div align="center">
37
-
38
- <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1200 360" width="100%" height="auto">
39
- <defs>
40
- <linearGradient id="bg" x1="0%" y1="0%" x2="100%" y2="100%">
41
- <stop offset="0%" stop-color="#050507"/>
42
- <stop offset="45%" stop-color="#0d0914"/>
43
- <stop offset="100%" stop-color="#071116"/>
44
- </linearGradient>
45
-
46
- <linearGradient id="obsidian" x1="0%" y1="0%" x2="100%" y2="0%">
47
- <stop offset="0%" stop-color="#7c3aed"/>
48
- <stop offset="45%" stop-color="#c026d3"/>
49
- <stop offset="100%" stop-color="#06b6d4"/>
50
- </linearGradient>
51
-
52
- <filter id="glow">
53
- <feGaussianBlur stdDeviation="5" result="blur"/>
54
- <feMerge>
55
- <feMergeNode in="blur"/>
56
- <feMergeNode in="SourceGraphic"/>
57
- </feMerge>
58
- </filter>
59
- </defs>
60
-
61
- <rect width="1200" height="360" rx="22" fill="url(#bg)"/>
62
-
63
- <rect x="2" y="2" width="1196" height="356" rx="20"
64
- fill="none"
65
- stroke="url(#obsidian)"
66
- stroke-width="2"
67
- opacity="0.5"/>
68
-
69
- <g opacity="0.07" stroke="#a855f7">
70
- <path d="M0 90H1200 M0 180H1200 M0 270H1200"/>
71
- <path d="M150 0V360 M300 0V360 M450 0V360 M600 0V360 M750 0V360 M900 0V360 M1050 0V360"/>
72
- </g>
73
-
74
- <g transform="translate(85,70)" filter="url(#glow)">
75
- <path d="M100 0 L180 62 L155 185 L100 225 L45 185 L20 62 Z"
76
- fill="#0c0912"
77
- stroke="url(#obsidian)"
78
- stroke-width="3"/>
79
-
80
- <path d="M100 25 L145 75 L128 160 L100 195 L72 160 L55 75 Z"
81
- fill="#171020"
82
- stroke="#8b5cf6"
83
- stroke-width="2"/>
84
-
85
- <path d="M100 45 L120 100 L100 170 L80 100 Z"
86
- fill="url(#obsidian)"/>
87
- </g>
88
-
89
- <text x="330" y="125"
90
- font-family="-apple-system,BlinkMacSystemFont,'Segoe UI',sans-serif"
91
- font-size="58"
92
- font-weight="900"
93
- fill="url(#obsidian)"
94
- letter-spacing="3">
95
- OBSIDIAN
96
- </text>
97
-
98
- <text x="334" y="177"
99
- font-family="-apple-system,BlinkMacSystemFont,'Segoe UI',sans-serif"
100
- font-size="30"
101
- font-weight="800"
102
- fill="#e2e8f0"
103
- letter-spacing="6">
104
- 9B CODER
105
- </text>
106
-
107
- <text x="334" y="218"
108
- font-family="monospace"
109
- font-size="16"
110
- fill="#94a3b8">
111
- QWEN CODER • MULTILINGUAL • FRONTEND SPECIALIZED
112
- </text>
113
-
114
- <g transform="translate(334,250)">
115
- <rect width="130" height="30" rx="15"
116
- fill="#7c3aed"
117
- fill-opacity="0.18"/>
118
- <text x="65" y="20"
119
- text-anchor="middle"
120
- font-family="monospace"
121
- font-size="12"
122
- fill="#c4b5fd">
123
- THREE.JS
124
- </text>
125
-
126
- <rect x="145" width="130" height="30" rx="15"
127
- fill="#c026d3"
128
- fill-opacity="0.16"/>
129
- <text x="210" y="20"
130
- text-anchor="middle"
131
- font-family="monospace"
132
- font-size="12"
133
- fill="#f0abfc">
134
- JAVASCRIPT
135
- </text>
136
-
137
- <rect x="290" width="130" height="30" rx="15"
138
- fill="#2563eb"
139
- fill-opacity="0.16"/>
140
- <text x="355" y="20"
141
- text-anchor="middle"
142
- font-family="monospace"
143
- font-size="12"
144
- fill="#93c5fd">
145
- HTML/CSS
146
- </text>
147
-
148
- <rect x="435" width="130" height="30" rx="15"
149
- fill="#0891b2"
150
- fill-opacity="0.16"/>
151
- <text x="500" y="20"
152
- text-anchor="middle"
153
- font-family="monospace"
154
- font-size="12"
155
- fill="#67e8f9">
156
- CANVAS
157
- </text>
158
- </g>
159
-
160
- <text x="334" y="320"
161
- font-family="monospace"
162
- font-size="13"
163
- fill="#64748b">
164
- FORGED FOR COMPLETE CODE GENERATION
165
- </text>
166
- </svg>
167
-
168
- </div>
169
-
170
  # OBSIDIAN-9B-Coder
171
 
172
  **OBSIDIAN-9B-Coder** is a 9B coding model fine-tuned on the **Coder Max Multilingual** dataset.
@@ -177,62 +34,22 @@ The model is specialized in generating **complete software implementations**, wi
177
 
178
  ---
179
 
180
- <div align="center">
181
-
182
- <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1200 125" width="100%">
183
- <rect width="1200" height="125" rx="18"
184
- fill="#08080d"
185
- stroke="#312e81"
186
- stroke-opacity="0.7"/>
187
-
188
- <g font-family="monospace" text-anchor="middle">
189
-
190
- <text x="120" y="48"
191
- font-size="27"
192
- font-weight="900"
193
- fill="#c4b5fd">9B</text>
194
- <text x="120" y="78"
195
- font-size="11"
196
- fill="#64748b">MODEL CLASS</text>
197
-
198
- <text x="360" y="48"
199
- font-size="27"
200
- font-weight="900"
201
- fill="#67e8f9">32K</text>
202
- <text x="360" y="78"
203
- font-size="11"
204
- fill="#64748b">TRAINING CONTEXT</text>
205
-
206
- <text x="600" y="48"
207
- font-size="27"
208
- font-weight="900"
209
- fill="#f0abfc">10</text>
210
- <text x="600" y="78"
211
- font-size="11"
212
- fill="#64748b">LANGUAGES</text>
213
-
214
- <text x="840" y="48"
215
- font-size="27"
216
- font-weight="900"
217
- fill="#93c5fd">LoRA</text>
218
- <text x="840" y="78"
219
- font-size="11"
220
- fill="#64748b">SFT</text>
221
-
222
- <text x="1080" y="48"
223
- font-size="27"
224
- font-weight="900"
225
- fill="#86efac">GGUF</text>
226
- <text x="1080" y="78"
227
- font-size="11"
228
- fill="#64748b">READY</text>
229
-
230
- </g>
231
- </svg>
232
-
233
- </div>
234
-
235
- ## Overview
236
 
237
  OBSIDIAN-9B-Coder was created to specialize a Qwen-based coding model toward **implementation-heavy programming tasks**.
238
 
@@ -240,128 +57,48 @@ Instead of focusing primarily on explanations surrounding code, the fine-tuning
240
 
241
  The model is particularly suited for:
242
 
243
- - Complete single-file web applications
244
- - HTML5 and modern CSS
245
- - JavaScript ES6+
246
- - Three.js
247
- - HTML5 Canvas
248
- - Interactive browser applications
249
- - Games and graphical experiments
250
- - DOM manipulation
251
- - Animation loops
252
- - State management
253
- - Python
254
- - Algorithmic programming
255
- - Multilingual programming instructions
256
 
257
  ---
258
 
259
- ## Core Philosophy
260
-
261
- <div align="center">
262
-
263
- <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1200 220" width="100%">
264
- <defs>
265
- <linearGradient id="flow" x1="0%" y1="0%" x2="100%" y2="0%">
266
- <stop offset="0%" stop-color="#7c3aed"/>
267
- <stop offset="50%" stop-color="#c026d3"/>
268
- <stop offset="100%" stop-color="#0891b2"/>
269
- </linearGradient>
270
- </defs>
271
-
272
- <rect width="1200" height="220" rx="18"
273
- fill="#07070b"
274
- stroke="#27272a"/>
275
-
276
- <text x="600" y="40"
277
- text-anchor="middle"
278
- font-family="monospace"
279
- font-size="15"
280
- font-weight="700"
281
- fill="#c4b5fd">
282
- FROM PROMPT TO COMPLETE SOFTWARE
283
- </text>
284
-
285
- <g font-family="monospace" text-anchor="middle">
286
-
287
- <rect x="75" y="80" width="230" height="70" rx="14"
288
- fill="#120d1b"
289
- stroke="#7c3aed"/>
290
-
291
- <text x="190" y="109"
292
- font-size="14"
293
- fill="#ddd6fe">
294
- USER REQUEST
295
- </text>
296
-
297
- <text x="190" y="132"
298
- font-size="11"
299
- fill="#64748b">
300
- natural language
301
- </text>
302
-
303
- <path d="M305 115H445"
304
- stroke="url(#flow)"
305
- stroke-width="4"/>
306
-
307
- <path d="M432 105L445 115L432 125"
308
- fill="none"
309
- stroke="#c026d3"
310
- stroke-width="4"/>
311
-
312
- <rect x="445" y="80" width="310" height="70" rx="14"
313
- fill="#150d18"
314
- stroke="#c026d3"/>
315
-
316
- <text x="600" y="109"
317
- font-size="14"
318
- fill="#f5d0fe">
319
- OBSIDIAN-9B-CODER
320
- </text>
321
-
322
- <text x="600" y="132"
323
- font-size="11"
324
- fill="#64748b">
325
- generate + construct
326
- </text>
327
-
328
- <path d="M755 115H895"
329
- stroke="url(#flow)"
330
- stroke-width="4"/>
331
-
332
- <path d="M882 105L895 115L882 125"
333
- fill="none"
334
- stroke="#0891b2"
335
- stroke-width="4"/>
336
-
337
- <rect x="895" y="80" width="230" height="70" rx="14"
338
- fill="#07171b"
339
- stroke="#0891b2"/>
340
-
341
- <text x="1010" y="109"
342
- font-size="14"
343
- fill="#a5f3fc">
344
- COMPLETE CODE
345
- </text>
346
-
347
- <text x="1010" y="132"
348
- font-size="11"
349
- fill="#64748b">
350
- full implementation
351
- </text>
352
-
353
- </g>
354
-
355
- <text x="600" y="190"
356
- text-anchor="middle"
357
- font-family="monospace"
358
- font-size="11"
359
- fill="#52525b">
360
- LESS BOILERPLATE • MORE IMPLEMENTATION • COMPLETE OUTPUTS
361
- </text>
362
- </svg>
363
-
364
- </div>
365
 
366
  ---
367
 
@@ -373,18 +110,18 @@ The repository contains multiple quantizations so users can select the appropria
373
 
374
  ## Available Model Files
375
 
376
- | File | Quantization | Notes |
377
- |---|---|---|
378
- | `Qwopus3.5-9B-Coder.Q8_0.gguf` | Q8_0 | Highest-quality quantized version |
379
- | `Qwopus3.5-9B-Coder.Q6_K.gguf` | Q6_K | High quality with reduced size |
380
- | `Qwopus3.5-9B-Coder.Q5_K_M.gguf` | Q5_K_M | Strong quality/size balance |
381
- | `Qwopus3.5-9B-Coder.Q4_K_M.gguf` | Q4_K_M | Recommended balance for local use |
382
- | `Qwopus3.5-9B-Coder.Q3_K_M.gguf` | Q3_K_M | Lower memory requirement |
383
- | `Qwopus3.5-9B-Coder.BF16-mmproj.gguf` | BF16 mmproj | Multimodal projector artifact |
384
 
385
  ---
386
 
387
- ## llama.cpp
388
 
389
  For compatible text-generation configurations:
390
 
@@ -402,67 +139,37 @@ Depending on your environment and the files selected automatically by `llama.cpp
402
 
403
  ---
404
 
405
- ## Quantization Guide
406
 
407
- <div align="center">
408
 
409
- <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1200 310" width="100%">
410
- <rect width="1200" height="310" rx="20"
411
- fill="#08080d"
412
- stroke="#27272a"/>
413
 
414
- <text x="600" y="42"
415
- text-anchor="middle"
416
- font-family="monospace"
417
- font-size="16"
418
- font-weight="700"
419
- fill="#c4b5fd">
420
- GGUF QUANTIZATION LADDER
421
- </text>
422
 
423
- <g font-family="monospace">
424
 
425
- <text x="90" y="92" fill="#94a3b8">Q8_0</text>
426
- <rect x="200" y="75" width="780" height="20" rx="10" fill="#18181b"/>
427
- <rect x="200" y="75" width="760" height="20" rx="10" fill="#7c3aed"/>
428
- <text x="1010" y="91" fill="#c4b5fd">QUALITY</text>
429
 
430
- <text x="90" y="132" fill="#94a3b8">Q6_K</text>
431
- <rect x="200" y="115" width="780" height="20" rx="10" fill="#18181b"/>
432
- <rect x="200" y="115" width="700" height="20" rx="10" fill="#8b5cf6"/>
433
- <text x="1010" y="131" fill="#c4b5fd">HIGH</text>
434
 
435
- <text x="90" y="172" fill="#94a3b8">Q5_K_M</text>
436
- <rect x="200" y="155" width="780" height="20" rx="10" fill="#18181b"/>
437
- <rect x="200" y="155" width="640" height="20" rx="10" fill="#c026d3"/>
438
- <text x="1010" y="171" fill="#f0abfc">BALANCED+</text>
439
 
440
- <text x="90" y="212" fill="#94a3b8">Q4_K_M</text>
441
- <rect x="200" y="195" width="780" height="20" rx="10" fill="#18181b"/>
442
- <rect x="200" y="195" width="570" height="20" rx="10" fill="#2563eb"/>
443
- <text x="1010" y="211" fill="#93c5fd">BALANCED</text>
444
 
445
- <text x="90" y="252" fill="#94a3b8">Q3_K_M</text>
446
- <rect x="200" y="235" width="780" height="20" rx="10" fill="#18181b"/>
447
- <rect x="200" y="235" width="470" height="20" rx="10" fill="#0891b2"/>
448
- <text x="1010" y="251" fill="#67e8f9">COMPACT</text>
449
 
450
- </g>
451
 
452
- <text x="600" y="288"
453
- text-anchor="middle"
454
- font-family="monospace"
455
- font-size="11"
456
- fill="#52525b">
457
- MORE BITS → MORE MEMORY • FEWER BITS → SMALLER FOOTPRINT
458
- </text>
459
- </svg>
460
 
461
- </div>
462
 
463
- For most local deployments, **Q4_K_M** or **Q5_K_M** are reasonable starting points.
464
 
465
- Use **Q8_0** when memory is less constrained and preserving model quality is more important.
466
 
467
  ---
468
 
@@ -470,43 +177,46 @@ Use **Q8_0** when memory is less constrained and preserving model quality is mor
470
 
471
  OBSIDIAN-9B-Coder was fine-tuned using:
472
 
473
- **Coder Max Multilingual**
 
 
 
474
  `guell00/Coder-max`
475
 
476
  Coder Max is a conversational supervised fine-tuning dataset focused heavily on code generation.
477
 
478
- The corpus was designed to contain complete implementations instead of heavily truncated programming answers.
479
 
480
  ## Dataset Characteristics
481
 
482
- | Characteristic | Description |
483
- |---|---|
484
- | Format | JSONL |
485
- | Structure | Conversational messages |
486
- | Training type | Supervised Fine-Tuning |
487
- | Languages | 10 |
488
- | Main content | Programming |
489
- | Code density | ~95%+ |
490
- | Frontend focus | Strong |
491
- | Three.js specialization | Strong |
492
- | Long code outputs | Preserved |
493
 
494
  ---
495
 
496
- ## Dataset Scale
497
 
498
  Coder Max is distributed in incremental variants.
499
 
500
- | Variant | Physical Size | Records | Messages | Code Density |
501
- |---|---:|---:|---:|---:|
502
- | `001MB` | 3,739,874 B | 109 | 220 | 99.08% |
503
- | `010MB` | 12,709,008 B | 969 | 2,012 | 96.18% |
504
- | `100MB` | 102,679,666 B | 9,790 | 20,396 | 95.84% |
505
- | `300MB` | 302,689,973 B | 29,233 | 60,916 | 95.83% |
506
- | `500MB` | 502,678,782 B | 48,676 | 101,442 | 95.82% |
507
- | `600MB` | 602,666,385 B | 58,466 | 121,848 | 95.82% |
508
- | `001GB` | 1,002,677,454 B | 97,499 | 203,200 | 95.82% |
509
- | `total_4GB` | 4,002,669,404 B | 390,302 | 813,452 | 95.81% |
510
 
511
  The larger variants contain the content represented by the smaller variants, allowing different training scales without requiring manual concatenation.
512
 
@@ -514,75 +224,27 @@ The larger variants contain the content represented by the smaller variants, all
514
 
515
  # Specialization
516
 
517
- <div align="center">
518
-
519
- <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1200 335" width="100%">
520
- <rect width="1200" height="335" rx="20"
521
- fill="#08080d"
522
- stroke="#27272a"/>
523
-
524
- <text x="600" y="43"
525
- text-anchor="middle"
526
- font-family="monospace"
527
- font-size="16"
528
- font-weight="700"
529
- fill="#e2e8f0">
530
- OBSIDIAN SPECIALIZATION
531
- </text>
532
-
533
- <g font-family="monospace" font-size="13">
534
-
535
- <text x="85" y="95" fill="#94a3b8">Three.js</text>
536
- <rect x="245" y="78" width="780" height="22" rx="11" fill="#18181b"/>
537
- <rect x="245" y="78" width="755" height="22" rx="11" fill="#7c3aed"/>
538
- <text x="1060" y="95" fill="#c4b5fd">CORE</text>
539
-
540
- <text x="85" y="140" fill="#94a3b8">JavaScript</text>
541
- <rect x="245" y="123" width="780" height="22" rx="11" fill="#18181b"/>
542
- <rect x="245" y="123" width="740" height="22" rx="11" fill="#a21caf"/>
543
- <text x="1060" y="140" fill="#f0abfc">HIGH</text>
544
-
545
- <text x="85" y="185" fill="#94a3b8">HTML / CSS</text>
546
- <rect x="245" y="168" width="780" height="22" rx="11" fill="#18181b"/>
547
- <rect x="245" y="168" width="720" height="22" rx="11" fill="#2563eb"/>
548
- <text x="1060" y="185" fill="#93c5fd">HIGH</text>
549
-
550
- <text x="85" y="230" fill="#94a3b8">Canvas</text>
551
- <rect x="245" y="213" width="780" height="22" rx="11" fill="#18181b"/>
552
- <rect x="245" y="213" width="680" height="22" rx="11" fill="#0891b2"/>
553
- <text x="1060" y="230" fill="#67e8f9">HIGH</text>
554
-
555
- <text x="85" y="275" fill="#94a3b8">Python</text>
556
- <rect x="245" y="258" width="780" height="22" rx="11" fill="#18181b"/>
557
- <rect x="245" y="258" width="610" height="22" rx="11" fill="#4f46e5"/>
558
- <text x="1060" y="275" fill="#a5b4fc">STRONG</text>
559
-
560
- </g>
561
- </svg>
562
-
563
- </div>
564
-
565
  ## Three.js
566
 
567
  Three.js is one of the primary specialization targets of OBSIDIAN.
568
 
569
  Training examples include patterns involving:
570
 
571
- - Scene creation
572
- - Perspective cameras
573
- - Lighting
574
- - Meshes
575
- - Materials
576
- - Geometry
577
- - Animation loops
578
- - Browser interaction
579
- - Keyboard input
580
- - Game mechanics
581
- - 3D environments
582
- - Interactive simulations
583
- - Single-file applications
584
-
585
- The goal is to teach the model to connect all required pieces of a Three.js application rather than merely outputting isolated API examples.
586
 
587
  ---
588
 
@@ -636,103 +298,22 @@ A typical training target may contain an entire application:
636
 
637
  # Multilingual Programming
638
 
639
- <div align="center">
640
-
641
- <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1200 180" width="100%">
642
- <rect width="1200" height="180" rx="18"
643
- fill="#09090e"
644
- stroke="#27272a"/>
645
-
646
- <text x="600" y="38"
647
- text-anchor="middle"
648
- font-family="monospace"
649
- font-size="15"
650
- font-weight="700"
651
- fill="#a78bfa">
652
- MULTILINGUAL CODE INSTRUCTIONS
653
- </text>
654
-
655
- <g font-family="monospace"
656
- font-size="14"
657
- text-anchor="middle">
658
-
659
- <g transform="translate(55,70)">
660
- <rect width="100" height="42" rx="10" fill="#18121f" stroke="#7c3aed"/>
661
- <text x="50" y="26" fill="#ddd6fe">PT</text>
662
- </g>
663
-
664
- <g transform="translate(165,70)">
665
- <rect width="100" height="42" rx="10" fill="#18121f" stroke="#8b5cf6"/>
666
- <text x="50" y="26" fill="#ddd6fe">EN</text>
667
- </g>
668
-
669
- <g transform="translate(275,70)">
670
- <rect width="100" height="42" rx="10" fill="#18121f" stroke="#a855f7"/>
671
- <text x="50" y="26" fill="#e9d5ff">ES</text>
672
- </g>
673
-
674
- <g transform="translate(385,70)">
675
- <rect width="100" height="42" rx="10" fill="#19111e" stroke="#c026d3"/>
676
- <text x="50" y="26" fill="#f5d0fe">FR</text>
677
- </g>
678
-
679
- <g transform="translate(495,70)">
680
- <rect width="100" height="42" rx="10" fill="#19111e" stroke="#db2777"/>
681
- <text x="50" y="26" fill="#fbcfe8">DE</text>
682
- </g>
683
-
684
- <g transform="translate(605,70)">
685
- <rect width="100" height="42" rx="10" fill="#10131f" stroke="#2563eb"/>
686
- <text x="50" y="26" fill="#bfdbfe">IT</text>
687
- </g>
688
-
689
- <g transform="translate(715,70)">
690
- <rect width="100" height="42" rx="10" fill="#0c151c" stroke="#0284c7"/>
691
- <text x="50" y="26" fill="#bae6fd">JA</text>
692
- </g>
693
-
694
- <g transform="translate(825,70)">
695
- <rect width="100" height="42" rx="10" fill="#0c1519" stroke="#0891b2"/>
696
- <text x="50" y="26" fill="#a5f3fc">ZH</text>
697
- </g>
698
-
699
- <g transform="translate(935,70)">
700
- <rect width="100" height="42" rx="10" fill="#0c1718" stroke="#0d9488"/>
701
- <text x="50" y="26" fill="#99f6e4">RU</text>
702
- </g>
703
-
704
- <g transform="translate(1045,70)">
705
- <rect width="100" height="42" rx="10" fill="#0c1715" stroke="#10b981"/>
706
- <text x="50" y="26" fill="#a7f3d0">TR</text>
707
- </g>
708
-
709
- </g>
710
-
711
- <text x="600" y="148"
712
- text-anchor="middle"
713
- font-family="monospace"
714
- font-size="11"
715
- fill="#52525b">
716
- 10-LANGUAGE PROMPT MATRIX
717
- </text>
718
- </svg>
719
-
720
- </div>
721
-
722
- The training corpus contains programming instructions across:
723
-
724
- | Language | Code |
725
- |---|---|
726
  | Português | `pt` |
727
- | English | `en` |
728
- | Español | `es` |
729
- | Français | `fr` |
730
- | Deutsch | `de` |
731
- | Italiano | `it` |
732
- | 日本語 | `ja` |
733
- | 简体中文 | `zh` |
734
- | Русский | `ru` |
735
- | Türkçe | `tr` |
 
 
736
 
737
  ---
738
 
@@ -771,9 +352,29 @@ down_proj
771
 
772
  ---
773
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
774
  # Example Prompts
775
 
776
- ### Three.js
777
 
778
  ```text
779
  Create a complete Three.js game inside a single HTML file.
@@ -791,7 +392,7 @@ Include:
791
  Return the complete HTML file.
792
  ```
793
 
794
- ### Portuguese
795
 
796
  ```text
797
  Crie uma aplicação web completa usando HTML, CSS e JavaScript.
@@ -802,7 +403,7 @@ estado interno e interação com o usuário.
802
  Retorne o arquivo HTML completo.
803
  ```
804
 
805
- ### Canvas
806
 
807
  ```text
808
  Build a complete interactive particle simulation using the
@@ -812,6 +413,14 @@ Include mouse interaction, animation, responsive resizing
812
  and performance-conscious rendering.
813
  ```
814
 
 
 
 
 
 
 
 
 
815
  ---
816
 
817
  # Recommended Generation
@@ -832,27 +441,26 @@ temperature: 0.5 - 0.7
832
  top_p: 0.9 - 0.95
833
  ```
834
 
835
- These values are starting points rather than guarantees. Generation settings should be benchmarked for the target workload.
836
-
837
- ---
838
 
839
- # Limitations
840
 
841
- OBSIDIAN-9B-Coder is a generative model.
842
 
843
- Generated code can contain:
844
 
845
- - logical errors;
846
- - security vulnerabilities;
847
- - hallucinated APIs;
848
- - outdated library usage;
849
- - incomplete edge-case handling;
850
- - incorrect assumptions;
851
- - dependency incompatibilities.
852
 
853
- Generated applications should be inspected and tested before production deployment.
854
-
855
- The model's context capacity also does not guarantee perfect reasoning or perfect retention across every token in a long context.
 
 
 
 
 
 
 
856
 
857
  ---
858
 
@@ -860,85 +468,55 @@ The model's context capacity also does not guarantee perfect reasoning or perfec
860
 
861
  For coding models such as OBSIDIAN, executable evaluation is strongly recommended.
862
 
863
- <div align="center">
864
-
865
- <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1200 210" width="100%">
866
- <rect width="1200" height="210" rx="18"
867
- fill="#08080c"
868
- stroke="#27272a"/>
869
-
870
- <text x="600" y="38"
871
- text-anchor="middle"
872
- font-family="monospace"
873
- font-size="15"
874
- font-weight="700"
875
- fill="#c4b5fd">
876
- CODE EVALUATION PIPELINE
877
- </text>
878
-
879
- <g font-family="monospace"
880
- font-size="12"
881
- text-anchor="middle">
882
-
883
- <rect x="50" y="85" width="170" height="55" rx="12"
884
- fill="#130d1c"
885
- stroke="#7c3aed"/>
886
- <text x="135" y="118" fill="#ddd6fe">GENERATE</text>
887
-
888
- <text x="250" y="118" fill="#7c3aed">→</text>
889
-
890
- <rect x="280" y="85" width="170" height="55" rx="12"
891
- fill="#160d18"
892
- stroke="#c026d3"/>
893
- <text x="365" y="118" fill="#f5d0fe">PARSE</text>
894
 
895
- <text x="480" y="118" fill="#c026d3">→</text>
 
 
 
 
 
 
 
 
 
 
896
 
897
- <rect x="510" y="85" width="170" height="55" rx="12"
898
- fill="#10121b"
899
- stroke="#2563eb"/>
900
- <text x="595" y="118" fill="#bfdbfe">EXECUTE</text>
901
 
902
- <text x="710" y="118" fill="#2563eb">→</text>
 
 
 
 
 
 
 
 
 
 
903
 
904
- <rect x="740" y="85" width="170" height="55" rx="12"
905
- fill="#08151a"
906
- stroke="#0891b2"/>
907
- <text x="825" y="118" fill="#a5f3fc">INSPECT</text>
908
 
909
- <text x="940" y="118" fill="#0891b2">→</text>
910
 
911
- <rect x="970" y="85" width="180" height="55" rx="12"
912
- fill="#091712"
913
- stroke="#10b981"/>
914
- <text x="1060" y="118" fill="#a7f3d0">TEST</text>
915
 
916
- </g>
917
 
918
- <text x="600" y="177"
919
- text-anchor="middle"
920
- font-family="monospace"
921
- font-size="11"
922
- fill="#52525b">
923
- EXECUTION MATTERS MORE THAN TEXT SIMILARITY
924
- </text>
925
- </svg>
926
 
927
- </div>
 
 
 
 
 
 
928
 
929
- Useful evaluation categories include:
930
 
931
- - HTML completeness
932
- - CSS validity
933
- - JavaScript syntax
934
- - Three.js initialization
935
- - Rendering-loop correctness
936
- - DOM interaction
937
- - Canvas rendering
938
- - Python syntax
939
- - Algorithmic correctness
940
- - Long-response completion
941
- - Multilingual instruction following
942
 
943
  ---
944
 
@@ -948,14 +526,12 @@ Coder Max includes metadata intended to support corpus integrity and provenance
948
 
949
  Dataset preparation includes:
950
 
951
- - syntax-oriented filtering;
952
- - preservation of complete code;
953
- - credential and runtime artifact sanitization;
954
- - SHA-256 provenance metadata;
955
- - multilingual prompt labeling;
956
- - frontend classification metadata.
957
-
958
- Some repetitions in the dataset may be intentional and used to reinforce selected programming patterns and technologies.
959
 
960
  ---
961
 
@@ -971,69 +547,47 @@ Users redistributing or deploying the model should review the licensing and usag
971
 
972
  ---
973
 
974
- <div align="center">
975
 
976
- <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1200 300" width="100%">
977
- <defs>
978
- <radialGradient id="finalGlow">
979
- <stop offset="0%" stop-color="#8b5cf6" stop-opacity="0.45"/>
980
- <stop offset="100%" stop-color="#8b5cf6" stop-opacity="0"/>
981
- </radialGradient>
982
 
983
- <linearGradient id="finalBlade" x1="0%" y1="0%" x2="100%" y2="100%">
984
- <stop offset="0%" stop-color="#c4b5fd"/>
985
- <stop offset="50%" stop-color="#8b5cf6"/>
986
- <stop offset="100%" stop-color="#22d3ee"/>
987
- </linearGradient>
988
- </defs>
989
 
990
- <rect width="1200" height="300" rx="20"
991
- fill="#060609"/>
992
 
993
- <circle cx="600" cy="135"
994
- r="130"
995
- fill="url(#finalGlow)"/>
996
 
997
- <path d="M600 25 L685 80 L655 195 L600 245 L545 195 L515 80 Z"
998
- fill="#0d0913"
999
- stroke="url(#finalBlade)"
1000
- stroke-width="3"/>
1001
 
1002
- <path d="M600 50 L645 95 L625 180 L600 215 L575 180 L555 95 Z"
1003
- fill="#151020"
1004
- stroke="#22d3ee"
1005
- stroke-opacity="0.6"/>
1006
 
1007
- <path d="M600 70 L619 118 L600 190 L581 118 Z"
1008
- fill="url(#finalBlade)"/>
1009
 
1010
- <text x="600" y="278"
1011
- text-anchor="middle"
1012
- font-family="monospace"
1013
- font-size="13"
1014
- font-weight="700"
1015
- fill="#64748b">
1016
- OBSIDIAN // FORGED FOR CODE
1017
- </text>
1018
- </svg>
1019
 
1020
- </div>
1021
 
1022
- # Author
1023
 
1024
- **guell00**
1025
 
1026
- **Model:** OBSIDIAN-9B-Coder
1027
- **Dataset:** Coder Max Multilingual
1028
 
1029
- ---
 
 
 
 
 
 
 
 
 
1030
 
1031
- <div align="center">
1032
 
1033
  ## OBSIDIAN-9B-Coder
1034
 
1035
  **Complete Code • Long Context • Interactive Software**
1036
 
1037
  `Three.js` • `JavaScript` • `HTML` • `CSS` • `Canvas` • `Python` • `Multilingual`
1038
-
1039
- </div>
 
1
  ---
 
 
 
2
  language:
3
  - pt
4
  - en
 
14
  tags:
15
  - code
16
  - coding
 
 
 
17
  - gguf
18
  - unsloth
19
  - lora
 
20
  - html
 
21
  - javascript
22
  - threejs
 
23
  - python
24
  - multilingual
25
  ---
26
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
27
  # OBSIDIAN-9B-Coder
28
 
29
  **OBSIDIAN-9B-Coder** is a 9B coding model fine-tuned on the **Coder Max Multilingual** dataset.
 
34
 
35
  ---
36
 
37
+ ## Highlights
38
+
39
+ * **9B parameter-class coding model**
40
+ * **32K training context**
41
+ * **Multilingual coding support**
42
+ * **Three.js specialization**
43
+ * **Complete HTML/CSS/JavaScript applications**
44
+ * **Canvas and browser graphics**
45
+ * **Python and algorithmic code**
46
+ * **LoRA supervised fine-tuning**
47
+ * **GGUF quantizations available**
48
+ * **Fine-tuned with Unsloth**
49
+
50
+ ---
51
+
52
+ # Overview
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
53
 
54
  OBSIDIAN-9B-Coder was created to specialize a Qwen-based coding model toward **implementation-heavy programming tasks**.
55
 
 
57
 
58
  The model is particularly suited for:
59
 
60
+ * Complete single-file web applications
61
+ * HTML5 and modern CSS
62
+ * JavaScript ES6+
63
+ * Three.js
64
+ * HTML5 Canvas
65
+ * Interactive browser applications
66
+ * Games and graphical experiments
67
+ * DOM manipulation
68
+ * Animation loops
69
+ * State management
70
+ * Python
71
+ * Algorithmic programming
72
+ * Multilingual programming instructions
73
 
74
  ---
75
 
76
+ # Core Philosophy
77
+
78
+ ```text
79
+ USER REQUEST
80
+
81
+ OBSIDIAN-9B-Coder
82
+
83
+ COMPLETE IMPLEMENTATION
84
+
85
+ HTML / CSS / JavaScript / Three.js / Python
86
+ ```
87
+
88
+ The goal is simple:
89
+
90
+ > **Less explanation. More implementation. Complete outputs.**
91
+
92
+ OBSIDIAN is fine-tuned to preserve long application structure, including:
93
+
94
+ * document layout;
95
+ * event handlers;
96
+ * state;
97
+ * rendering loops;
98
+ * interaction logic;
99
+ * animation;
100
+ * closing tags;
101
+ * frontend lifecycle.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
102
 
103
  ---
104
 
 
110
 
111
  ## Available Model Files
112
 
113
+ | File | Quantization | Notes |
114
+ | ------------------------------------- | ------------ | --------------------------------- |
115
+ | `Qwopus3.5-9B-Coder.Q8_0.gguf` | Q8_0 | Highest-quality quantized version |
116
+ | `Qwopus3.5-9B-Coder.Q6_K.gguf` | Q6_K | High quality with reduced size |
117
+ | `Qwopus3.5-9B-Coder.Q5_K_M.gguf` | Q5_K_M | Strong quality/size balance |
118
+ | `Qwopus3.5-9B-Coder.Q4_K_M.gguf` | Q4_K_M | Recommended balance for local use |
119
+ | `Qwopus3.5-9B-Coder.Q3_K_M.gguf` | Q3_K_M | Lower memory requirement |
120
+ | `Qwopus3.5-9B-Coder.BF16-mmproj.gguf` | BF16 mmproj | Multimodal projector artifact |
121
 
122
  ---
123
 
124
+ # llama.cpp
125
 
126
  For compatible text-generation configurations:
127
 
 
139
 
140
  ---
141
 
142
+ # Quantization Guide
143
 
144
+ ## Q8_0
145
 
146
+ Best option when memory usage is less important and you want to preserve as much quality as practical in GGUF form.
 
 
 
147
 
148
+ ## Q6_K
 
 
 
 
 
 
 
149
 
150
+ High-quality quantization with lower memory requirements than Q8.
151
 
152
+ ## Q5_K_M
 
 
 
153
 
154
+ A strong balance between quality and size.
 
 
 
155
 
156
+ Good choice for machines with moderate RAM/VRAM.
 
 
 
157
 
158
+ ## Q4_K_M
 
 
 
159
 
160
+ Recommended general-purpose quantization.
 
 
 
161
 
162
+ It usually offers a useful balance between:
163
 
164
+ * memory usage;
165
+ * inference speed;
166
+ * output quality.
 
 
 
 
 
167
 
168
+ ## Q3_K_M
169
 
170
+ Smaller memory footprint.
171
 
172
+ Useful when hardware is constrained, at the cost of some model fidelity.
173
 
174
  ---
175
 
 
177
 
178
  OBSIDIAN-9B-Coder was fine-tuned using:
179
 
180
+ **Coder Max Multilingual**
181
+
182
+ Repository:
183
+
184
  `guell00/Coder-max`
185
 
186
  Coder Max is a conversational supervised fine-tuning dataset focused heavily on code generation.
187
 
188
+ The corpus was designed to contain **complete implementations instead of heavily truncated programming answers**.
189
 
190
  ## Dataset Characteristics
191
 
192
+ | Characteristic | Description |
193
+ | ----------------------- | ----------------------- |
194
+ | Format | JSONL |
195
+ | Structure | Conversational messages |
196
+ | Training type | Supervised Fine-Tuning |
197
+ | Languages | 10 |
198
+ | Main content | Programming |
199
+ | Code density | ~95%+ |
200
+ | Frontend focus | Strong |
201
+ | Three.js specialization | Strong |
202
+ | Long code outputs | Preserved |
203
 
204
  ---
205
 
206
+ # Dataset Scale
207
 
208
  Coder Max is distributed in incremental variants.
209
 
210
+ | Variant | Physical Size | Records | Messages | Code Density |
211
+ | ----------- | --------------: | ------: | -------: | -----------: |
212
+ | `001MB` | 3,739,874 B | 109 | 220 | 99.08% |
213
+ | `010MB` | 12,709,008 B | 969 | 2,012 | 96.18% |
214
+ | `100MB` | 102,679,666 B | 9,790 | 20,396 | 95.84% |
215
+ | `300MB` | 302,689,973 B | 29,233 | 60,916 | 95.83% |
216
+ | `500MB` | 502,678,782 B | 48,676 | 101,442 | 95.82% |
217
+ | `600MB` | 602,666,385 B | 58,466 | 121,848 | 95.82% |
218
+ | `001GB` | 1,002,677,454 B | 97,499 | 203,200 | 95.82% |
219
+ | `total_4GB` | 4,002,669,404 B | 390,302 | 813,452 | 95.81% |
220
 
221
  The larger variants contain the content represented by the smaller variants, allowing different training scales without requiring manual concatenation.
222
 
 
224
 
225
  # Specialization
226
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
227
  ## Three.js
228
 
229
  Three.js is one of the primary specialization targets of OBSIDIAN.
230
 
231
  Training examples include patterns involving:
232
 
233
+ * scene creation;
234
+ * perspective cameras;
235
+ * lighting;
236
+ * meshes;
237
+ * materials;
238
+ * geometry;
239
+ * animation loops;
240
+ * browser interaction;
241
+ * keyboard input;
242
+ * game mechanics;
243
+ * 3D environments;
244
+ * interactive simulations;
245
+ * single-file applications.
246
+
247
+ The goal is to teach the model to connect all required pieces of a Three.js application rather than merely output isolated API examples.
248
 
249
  ---
250
 
 
298
 
299
  # Multilingual Programming
300
 
301
+ The training corpus contains programming instructions across 10 languages.
302
+
303
+ | Language | Code |
304
+ | --------- | ---- |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
305
  | Português | `pt` |
306
+ | English | `en` |
307
+ | Español | `es` |
308
+ | Français | `fr` |
309
+ | Deutsch | `de` |
310
+ | Italiano | `it` |
311
+ | 日本語 | `ja` |
312
+ | 简体中文 | `zh` |
313
+ | Русский | `ru` |
314
+ | Türkçe | `tr` |
315
+
316
+ The objective is to make programming capability less dependent on the language used in the instruction.
317
 
318
  ---
319
 
 
352
 
353
  ---
354
 
355
+ # Training Strategy
356
+
357
+ The model was trained with a code-heavy SFT corpus designed around long-form completions.
358
+
359
+ Important characteristics include:
360
+
361
+ * long HTML responses;
362
+ * complete application generation;
363
+ * repeated high-value specialization examples;
364
+ * multilingual prompt variants;
365
+ * frontend-oriented training;
366
+ * Three.js and Canvas exposure;
367
+ * assistant-response-only loss.
368
+
369
+ Intentional repetitions may exist in the source dataset as part of the specialization strategy.
370
+
371
+ These repetitions are used to reinforce selected programming behaviors rather than representing accidental duplicate ingestion.
372
+
373
+ ---
374
+
375
  # Example Prompts
376
 
377
+ ## Three.js
378
 
379
  ```text
380
  Create a complete Three.js game inside a single HTML file.
 
392
  Return the complete HTML file.
393
  ```
394
 
395
+ ## Portuguese
396
 
397
  ```text
398
  Crie uma aplicação web completa usando HTML, CSS e JavaScript.
 
403
  Retorne o arquivo HTML completo.
404
  ```
405
 
406
+ ## Canvas
407
 
408
  ```text
409
  Build a complete interactive particle simulation using the
 
413
  and performance-conscious rendering.
414
  ```
415
 
416
+ ## Python
417
+
418
+ ```text
419
+ Implement a complete Python solution for the following problem.
420
+
421
+ Explain the algorithm briefly and return working code.
422
+ ```
423
+
424
  ---
425
 
426
  # Recommended Generation
 
441
  top_p: 0.9 - 0.95
442
  ```
443
 
444
+ These values are starting points rather than guarantees.
 
 
445
 
446
+ Generation settings should be benchmarked for the target workload.
447
 
448
+ ---
449
 
450
+ # Intended Use
451
 
452
+ OBSIDIAN-9B-Coder is intended for:
 
 
 
 
 
 
453
 
454
+ * coding assistants;
455
+ * frontend code generation;
456
+ * HTML/CSS/JavaScript generation;
457
+ * Three.js generation;
458
+ * Canvas applications;
459
+ * Python programming;
460
+ * programming experiments;
461
+ * multilingual coding assistants;
462
+ * local coding models;
463
+ * research into code-specialized fine-tuning.
464
 
465
  ---
466
 
 
468
 
469
  For coding models such as OBSIDIAN, executable evaluation is strongly recommended.
470
 
471
+ Useful evaluation categories include:
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
472
 
473
+ * HTML completeness;
474
+ * CSS validity;
475
+ * JavaScript syntax;
476
+ * Three.js initialization;
477
+ * rendering-loop correctness;
478
+ * DOM interaction;
479
+ * Canvas rendering;
480
+ * Python syntax;
481
+ * algorithmic correctness;
482
+ * long-response completion;
483
+ * multilingual instruction following.
484
 
485
+ A useful evaluation pipeline is:
 
 
 
486
 
487
+ ```text
488
+ GENERATE
489
+
490
+ PARSE
491
+
492
+ EXECUTE
493
+
494
+ INSPECT
495
+
496
+ TEST
497
+ ```
498
 
499
+ Execution matters more than text similarity when evaluating generated software.
 
 
 
500
 
501
+ ---
502
 
503
+ # Limitations
 
 
 
504
 
505
+ OBSIDIAN-9B-Coder is a generative model.
506
 
507
+ Generated code can contain:
 
 
 
 
 
 
 
508
 
509
+ * logical errors;
510
+ * security vulnerabilities;
511
+ * hallucinated APIs;
512
+ * outdated library usage;
513
+ * incomplete edge-case handling;
514
+ * incorrect assumptions;
515
+ * dependency incompatibilities.
516
 
517
+ Generated applications should be inspected and tested before production deployment.
518
 
519
+ The model's context capacity also does not guarantee perfect reasoning or perfect retention across every token in a long context.
 
 
 
 
 
 
 
 
 
 
520
 
521
  ---
522
 
 
526
 
527
  Dataset preparation includes:
528
 
529
+ * syntax-oriented filtering;
530
+ * preservation of complete code;
531
+ * credential and runtime artifact sanitization;
532
+ * SHA-256 provenance metadata;
533
+ * multilingual prompt labeling;
534
+ * frontend classification metadata.
 
 
535
 
536
  ---
537
 
 
547
 
548
  ---
549
 
550
+ # Model Identity
551
 
552
+ **Model:** OBSIDIAN-9B-Coder
 
 
 
 
 
553
 
554
+ **Dataset:** Coder Max Multilingual
 
 
 
 
 
555
 
556
+ **Author:** guell00
 
557
 
558
+ **Training framework:** Unsloth
 
 
559
 
560
+ **Format:** GGUF
 
 
 
561
 
562
+ **Specialization:**
 
 
 
563
 
564
+ `Three.js` `JavaScript` `HTML` `CSS` `Canvas` • `Python` • `Multilingual`
 
565
 
566
+ ---
 
 
 
 
 
 
 
 
567
 
568
+ # Summary
569
 
570
+ OBSIDIAN-9B-Coder is a 9B coding model specialized through supervised fine-tuning on a highly code-dense multilingual dataset.
571
 
572
+ Its main focus is:
573
 
574
+ > **Generating complete software instead of fragmented examples.**
 
575
 
576
+ Primary strengths include:
577
+
578
+ * frontend code generation;
579
+ * Three.js;
580
+ * JavaScript;
581
+ * complete HTML applications;
582
+ * Canvas;
583
+ * Python;
584
+ * multilingual programming instructions;
585
+ * long code outputs.
586
 
587
+ ---
588
 
589
  ## OBSIDIAN-9B-Coder
590
 
591
  **Complete Code • Long Context • Interactive Software**
592
 
593
  `Three.js` • `JavaScript` • `HTML` • `CSS` • `Canvas` • `Python` • `Multilingual`