Hugging Face's logo

Spaces:
hf-skills
/
llm-trainer
Running

App Files Community
burtenshaw HF Staff commited on
Commit
2444345
·
verified ·
1 Parent(s): 03bd9c8

Create index.html

Browse files
Files changed (1) hide show
  1. index.html +257 -0
index.html ADDED
@@ -0,0 +1,257 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <!DOCTYPE html>
2
+ <html lang="en">
3
+ <head>
4
+ <meta charset="UTF-8">
5
+ <meta name="viewport" content="width=device-width, initial-scale=1.0">
6
+ <title>Model Trainer Skill - Hugging Face</title>
7
+ <style>
8
+ * {
9
+ margin: 0;
10
+ padding: 0;
11
+ box-sizing: border-box;
12
+ }
13
+
14
+ body {
15
+ font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, 'Helvetica Neue', Arial, sans-serif;
16
+ line-height: 1.6;
17
+ color: #333;
18
+ background: #f5f5f5;
19
+ padding: 20px;
20
+ }
21
+
22
+ .container {
23
+ max-width: 1200px;
24
+ margin: 0 auto;
25
+ background: white;
26
+ padding: 40px;
27
+ border-radius: 8px;
28
+ box-shadow: 0 2px 4px rgba(0,0,0,0.1);
29
+ }
30
+
31
+ h1 {
32
+ color: #ffd21e;
33
+ background: #000;
34
+ padding: 20px;
35
+ margin: -40px -40px 30px -40px;
36
+ border-radius: 8px 8px 0 0;
37
+ }
38
+
39
+ h2 {
40
+ color: #1e1e1e;
41
+ margin-top: 30px;
42
+ margin-bottom: 15px;
43
+ padding-bottom: 10px;
44
+ border-bottom: 2px solid #ffd21e;
45
+ }
46
+
47
+ h3 {
48
+ color: #555;
49
+ margin-top: 20px;
50
+ margin-bottom: 10px;
51
+ }
52
+
53
+ .description {
54
+ background: #f9f9f9;
55
+ padding: 20px;
56
+ border-left: 4px solid #ffd21e;
57
+ margin-bottom: 30px;
58
+ border-radius: 4px;
59
+ }
60
+
61
+ .file-list {
62
+ list-style: none;
63
+ padding: 0;
64
+ }
65
+
66
+ .file-list li {
67
+ padding: 12px;
68
+ margin: 8px 0;
69
+ background: #f9f9f9;
70
+ border-radius: 4px;
71
+ border-left: 3px solid #ffd21e;
72
+ transition: background 0.2s;
73
+ }
74
+
75
+ .file-list li:hover {
76
+ background: #f0f0f0;
77
+ }
78
+
79
+ .file-list a {
80
+ color: #0066cc;
81
+ text-decoration: none;
82
+ font-weight: 500;
83
+ display: block;
84
+ }
85
+
86
+ .file-list a:hover {
87
+ text-decoration: underline;
88
+ }
89
+
90
+ .file-path {
91
+ color: #666;
92
+ font-size: 0.9em;
93
+ font-family: 'Monaco', 'Courier New', monospace;
94
+ margin-top: 4px;
95
+ }
96
+
97
+ .file-description {
98
+ color: #777;
99
+ font-size: 0.9em;
100
+ margin-top: 4px;
101
+ font-style: italic;
102
+ }
103
+
104
+ .metadata {
105
+ background: #f0f0f0;
106
+ padding: 15px;
107
+ border-radius: 4px;
108
+ margin-bottom: 30px;
109
+ }
110
+
111
+ .metadata p {
112
+ margin: 5px 0;
113
+ }
114
+
115
+ .metadata strong {
116
+ color: #333;
117
+ }
118
+
119
+ .section {
120
+ margin-bottom: 40px;
121
+ }
122
+
123
+ code {
124
+ background: #f4f4f4;
125
+ padding: 2px 6px;
126
+ border-radius: 3px;
127
+ font-family: 'Monaco', 'Courier New', monospace;
128
+ font-size: 0.9em;
129
+ }
130
+ </style>
131
+ </head>
132
+ <body>
133
+ <div class="container">
134
+ <h1>Agent Skill : LLM Training</h1>
135
+
136
+ <div class="description">
137
+ <p><strong>Train and fine-tune language models using TRL (Transformer Reinforcement Learning) on Hugging Face Jobs infrastructure.</strong></p>
138
+ <p>Covers SFT, DPO, GRPO and reward modeling training methods, plus GGUF conversion for local deployment. Includes guidance on the TRL Jobs package, UV scripts with PEP 723 format, dataset preparation and validation, hardware selection, cost estimation, Trackio monitoring, Hub authentication, and model persistence.</p>
139
+ </div>
140
+
141
+ <div class="metadata">
142
+ <p><strong>Skill Name:</strong> model-trainer</p>
143
+ <p><strong>Version:</strong> 1.0.0</p>
144
+ <p><strong>License:</strong> Apache-2.0</p>
145
+ <p><strong>Main Documentation:</strong> <a href="skills/model-trainer/SKILL.md">skills/model-trainer/SKILL.md</a></p>
146
+ </div>
147
+
148
+ <div class="section">
149
+ <h2>Core Documentation</h2>
150
+ <ul class="file-list">
151
+ <li>
152
+ <a href="skills/model-trainer/SKILL.md">SKILL.md</a>
153
+ <div class="file-path">skills/model-trainer/SKILL.md</div>
154
+ <div class="file-description">Complete skill documentation with training guides, examples, and best practices</div>
155
+ </li>
156
+ </ul>
157
+ </div>
158
+
159
+ <div class="section">
160
+ <h2>Training Scripts</h2>
161
+ <p>Production-ready training scripts with Trackio integration, LoRA/PEFT, and Hub saving:</p>
162
+ <ul class="file-list">
163
+ <li>
164
+ <a href="skills/model-trainer/scripts/train_sft_example.py">train_sft_example.py</a>
165
+ <div class="file-path">skills/model-trainer/scripts/train_sft_example.py</div>
166
+ <div class="file-description">Complete SFT (Supervised Fine-Tuning) training template with all best practices</div>
167
+ </li>
168
+ <li>
169
+ <a href="skills/model-trainer/scripts/train_dpo_example.py">train_dpo_example.py</a>
170
+ <div class="file-path">skills/model-trainer/scripts/train_dpo_example.py</div>
171
+ <div class="file-description">DPO (Direct Preference Optimization) training for preference learning</div>
172
+ </li>
173
+ <li>
174
+ <a href="skills/model-trainer/scripts/train_grpo_example.py">train_grpo_example.py</a>
175
+ <div class="file-path">skills/model-trainer/scripts/train_grpo_example.py</div>
176
+ <div class="file-description">GRPO (Group Relative Policy Optimization) training for online RL</div>
177
+ </li>
178
+ <li>
179
+ <a href="skills/model-trainer/scripts/convert_to_gguf.py">convert_to_gguf.py</a>
180
+ <div class="file-path">skills/model-trainer/scripts/convert_to_gguf.py</div>
181
+ <div class="file-description">Complete GGUF conversion script for local deployment (Ollama, LM Studio, llama.cpp)</div>
182
+ </li>
183
+ <li>
184
+ <a href="skills/model-trainer/scripts/estimate_cost.py">estimate_cost.py</a>
185
+ <div class="file-path">skills/model-trainer/scripts/estimate_cost.py</div>
186
+ <div class="file-description">Estimate training time and cost before launching jobs</div>
187
+ </li>
188
+ <li>
189
+ <a href="skills/model-trainer/scripts/dataset_inspector.py">dataset_inspector.py</a>
190
+ <div class="file-path">skills/model-trainer/scripts/dataset_inspector.py</div>
191
+ <div class="file-description">Validate dataset format before GPU training to prevent format failures</div>
192
+ </li>
193
+ </ul>
194
+ </div>
195
+
196
+ <div class="section">
197
+ <h2>Reference Guides</h2>
198
+ <p>Detailed guides for specific aspects of training:</p>
199
+ <ul class="file-list">
200
+ <li>
201
+ <a href="skills/model-trainer/references/training_methods.md">training_methods.md</a>
202
+ <div class="file-path">skills/model-trainer/references/training_methods.md</div>
203
+ <div class="file-description">Overview of SFT, DPO, GRPO, KTO, PPO, and Reward Modeling methods</div>
204
+ </li>
205
+ <li>
206
+ <a href="skills/model-trainer/references/training_patterns.md">training_patterns.md</a>
207
+ <div class="file-path">skills/model-trainer/references/training_patterns.md</div>
208
+ <div class="file-description">Common training patterns and examples (quick demo, production, multi-GPU)</div>
209
+ </li>
210
+ <li>
211
+ <a href="skills/model-trainer/references/hardware_guide.md">hardware_guide.md</a>
212
+ <div class="file-path">skills/model-trainer/references/hardware_guide.md</div>
213
+ <div class="file-description">Hardware specifications and selection guidelines for different model sizes</div>
214
+ </li>
215
+ <li>
216
+ <a href="skills/model-trainer/references/gguf_conversion.md">gguf_conversion.md</a>
217
+ <div class="file-path">skills/model-trainer/references/gguf_conversion.md</div>
218
+ <div class="file-description">Complete GGUF conversion guide with quantization options and usage examples</div>
219
+ </li>
220
+ <li>
221
+ <a href="skills/model-trainer/references/trackio_guide.md">trackio_guide.md</a>
222
+ <div class="file-path">skills/model-trainer/references/trackio_guide.md</div>
223
+ <div class="file-description">Trackio monitoring setup for real-time training metrics visualization</div>
224
+ </li>
225
+ <li>
226
+ <a href="skills/model-trainer/references/hub_saving.md">hub_saving.md</a>
227
+ <div class="file-path">skills/model-trainer/references/hub_saving.md</div>
228
+ <div class="file-description">Hub authentication troubleshooting and model persistence best practices</div>
229
+ </li>
230
+ <li>
231
+ <a href="skills/model-trainer/references/reliability_principles.md">reliability_principles.md</a>
232
+ <div class="file-path">skills/model-trainer/references/reliability_principles.md</div>
233
+ <div class="file-description">Reliability principles for production training workflows</div>
234
+ </li>
235
+ <li>
236
+ <a href="skills/model-trainer/references/troubleshooting.md">troubleshooting.md</a>
237
+ <div class="file-path">skills/model-trainer/references/troubleshooting.md</div>
238
+ <div class="file-description">Common issues and solutions (OOM, timeouts, Hub push failures, etc.)</div>
239
+ </li>
240
+ </ul>
241
+ </div>
242
+
243
+ <div class="section">
244
+ <h2>Plugin Configuration</h2>
245
+ <ul class="file-list">
246
+ <li>
247
+ <a href="plugin.json">plugin.json</a>
248
+ <div class="file-path">plugin.json</div>
249
+ <div class="file-description">Plugin metadata and configuration</div>
250
+ </li>
251
+ </ul>
252
+ </div>
253
+
254
+ </div>
255
+ </body>
256
+ </html>
257
+