Hugging Face's logo

Spaces:
nanotron
/
ultrascale-playbook
Running

App Files Community
126
thomwolf HF Staff commited on
Commit
9905b93
·
1 Parent(s): 6a71dad

update

Browse files
Files changed (5) hide show
  1. dist/index.html +128 -117
  2. dist/main.bundle.js +0 -0
  3. dist/main.bundle.js.map +0 -0
  4. src/index.html +128 -117
  5. src/memory.js +4 -4
dist/index.html CHANGED
@@ -9,32 +9,39 @@
9
  <title>FineWeb: decanting the web for the finest text data at scale</title>
10
  <link rel="stylesheet" href="style.css">
11
  <style>
12
- #controls {
13
- display: grid;
14
- grid-template-columns: auto 1fr auto;
15
- gap: 5px;
16
- align-items: center;
17
- max-width: 600px;
18
- margin-bottom: 20px;
19
- }
20
- #controls label {
21
- text-align: right;
22
- }
23
- #controls input[type="range"] {
24
- width: 100%;
25
- }
26
- #controls input[type="number"] {
27
- width: 60px;
28
- }
29
- #controls .row {
30
- display: contents;
31
- }
32
- #graph svg {
33
- font-family: sans-serif;
34
- }
35
- #graph svg rect {
36
- cursor: pointer;
37
- }
 
 
 
 
 
 
 
38
  </style>
39
  </head>
40
 
@@ -90,98 +97,102 @@
90
 
91
  <div id="graph" style="position: relative; width: 700px; height: 500px;"></div>
92
  <div id="controls">
93
- <div class="row">
94
- <label for="a">Attention Heads (a):</label>
95
- <input type="range" id="a" name="a" min="1" max="128" value="8">
96
- <input type="number" id="a_input" value="8" min="1" max="128">
97
- </div>
98
- <div class="row">
99
- <label for="b">Micro Batch Size (b):</label>
100
- <input type="range" id="b" name="b" min="1" max="53248" value="32">
101
- <input type="number" id="b_input" value="32" min="1" max="53248">
102
- </div>
103
- <div class="row">
104
- <label for="h">Hidden Dimension Size (h):</label>
105
- <input type="range" id="h" name="h" min="1" max="16384" value="512">
106
- <input type="number" id="h_input" value="512" min="128" max="16384">
107
- </div>
108
- <div class="row">
109
- <label for="h_ff">Feedforward Dimension Size (h_ff):</label>
110
- <input type="range" id="h_ff" name="h_ff" min="1" max="65536" value="2048">
111
- <input type="number" id="h_ff_input" value="2048" min="512" max="65536">
112
- </div>
113
- <div class="row">
114
- <label for="L">Number of Layers (L):</label>
115
- <input type="range" id="L" name="L" min="1" max="126" value="12">
116
- <input type="number" id="L_input" value="12" min="1" max="126">
117
- </div>
118
- <div class="row">
119
- <label for="s">Sequence Length (s):</label>
120
- <input type="range" id="s" name="s" min="1" max="128000" value="128">
121
- <input type="number" id="s_input" value="128" min="64" max="128000">
122
- </div>
123
- <div class="row">
124
- <label for="v">Vocabulary Size (v):</label>
125
- <input type="range" id="v" name="v" min="1000" max="100000" value="30522">
126
- <input type="number" id="v_input" value="30522" min="1000" max="100000">
127
- </div>
128
- <div class="row">
129
- <label for="k">Optimizer Parameters (k):</label>
130
- <input type="range" id="k" name="k" min="1" max="16" value="8">
131
- <input type="number" id="k_input" value="8" min="1" max="16">
132
- </div>
133
- <div class="row">
134
- <label for="tp">Tensor Model Parallelism (t):</label>
135
- <input type="range" id="tp" name="tp" min="1" max="16" value="8">
136
- <input type="number" id="tp_input" value="8" min="1" max="16">
137
- </div>
138
- <div class="row">
139
- <label for="dp">Data Model Parallelism (d):</label>
140
- <input type="range" id="dp" name="dp" min="1" max="256" value="1">
141
- <input type="number" id="dp_input" value="1" min="1" max="256">
142
- </div>
143
- <div class="row">
144
- <label for="mixed">Mixed Precision:</label>
145
- <input type="checkbox" id="mixed" name="mixed" checked>
146
- <span></span> <!-- Empty span to maintain grid alignment -->
147
- </div>
148
- <div class="row">
149
- <label for="recomputation">Recomputation:</label>
150
- <select id="recomputation" name="recomputation">
151
- <option value="none">None</option>
152
- <option value="selective">Selective</option>
153
- <option value="full">Full</option>
154
- </select>
155
- <span></span> <!-- Empty span to maintain grid alignment -->
156
- </div>
157
- <div class="row">
158
- <label for="zero">Zero:</label>
159
- <select id="zero" name="zero">
160
- <option value="0">0</option>
161
- <option value="1">1</option>
162
- <option value="2">2</option>
163
- <option value="3">3</option>
164
- </select>
165
- <span></span> <!-- Empty span to maintain grid alignment -->
166
- </div>
167
- <div class="row">
168
- <label for="ff_activation">FF Activation:</label>
169
- <select id="ff_activation" name="ff_activation">
170
- <option value="relu">ReLU</option>
171
- <option value="gelu">GELU</option>
172
- <option value="swiglu">SwiGLU</option>
173
- </select>
174
- <span></span> <!-- Empty span to maintain grid alignment -->
175
  </div>
176
- <div class="row">
177
- <label for="presets">Presets:</label>
178
- <select id="presets" name="presets">
179
- <option value="Llama 3 Tiny">Llama 3 Tiny</option>
180
- <option value="Llama 3 8B">Llama 3 8B</option>
181
- <option value="Llama 3 70B">Llama 3 70B</option>
182
- <option value="Llama 3 405B">Llama 3 405B</option>
183
- </select>
184
- <span></span> <!-- Empty span to maintain grid alignment -->
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
185
  </div>
186
  </div>
187
 
 
9
  <title>FineWeb: decanting the web for the finest text data at scale</title>
10
  <link rel="stylesheet" href="style.css">
11
  <style>
12
+ #controls {
13
+ display: grid;
14
+ grid-template-columns: auto 1fr auto;
15
+ gap: 5px;
16
+ align-items: center;
17
+ max-width: 700px;
18
+ margin-bottom: 10px;
19
+ }
20
+ #controls .row {
21
+ display: contents;
22
+ }
23
+ #controls label {
24
+ text-align: right;
25
+ padding-right: 10px;
26
+ }
27
+ #controls input[type="range"] {
28
+ width: 100%;
29
+ }
30
+ #controls input[type="number"] {
31
+ width: 60px;
32
+ }
33
+ #controls select {
34
+ width: 100%;
35
+ }
36
+ #controls .column {
37
+ display: contents;
38
+ }
39
+ #graph svg {
40
+ font-family: sans-serif;
41
+ }
42
+ #graph svg rect {
43
+ cursor: pointer;
44
+ }
45
  </style>
46
  </head>
47
 
 
97
 
98
  <div id="graph" style="position: relative; width: 700px; height: 500px;"></div>
99
  <div id="controls">
100
+ <div class="column">
101
+ <div class="row">
102
+ <label for="a">Attention Heads (a):</label>
103
+ <input type="range" id="a" name="a" min="1" max="128" value="8">
104
+ <input type="number" id="a_input" value="8" min="1" max="128">
105
+ </div>
106
+ <div class="row">
107
+ <label for="b">Micro Batch Size (b):</label>
108
+ <input type="range" id="b" name="b" min="1" max="53248" value="32">
109
+ <input type="number" id="b_input" value="32" min="1" max="53248">
110
+ </div>
111
+ <div class="row">
112
+ <label for="h">Hidden Dimension Size (h):</label>
113
+ <input type="range" id="h" name="h" min="1" max="16384" value="512">
114
+ <input type="number" id="h_input" value="512" min="128" max="16384">
115
+ </div>
116
+ <div class="row">
117
+ <label for="h_ff">Feedforward Dimension Size (h_ff):</label>
118
+ <input type="range" id="h_ff" name="h_ff" min="1" max="65536" value="2048">
119
+ <input type="number" id="h_ff_input" value="2048" min="512" max="65536">
120
+ </div>
121
+ <div class="row">
122
+ <label for="L">Number of Layers (L):</label>
123
+ <input type="range" id="L" name="L" min="1" max="126" value="12">
124
+ <input type="number" id="L_input" value="12" min="1" max="126">
125
+ </div>
126
+ <div class="row">
127
+ <label for="s">Sequence Length (s):</label>
128
+ <input type="range" id="s" name="s" min="1" max="128000" value="128">
129
+ <input type="number" id="s_input" value="128" min="64" max="128000">
130
+ </div>
131
+ <div class="row">
132
+ <label for="v">Vocabulary Size (v):</label>
133
+ <input type="range" id="v" name="v" min="1000" max="100000" value="30522">
134
+ <input type="number" id="v_input" value="30522" min="1000" max="100000">
135
+ </div>
136
+ <div class="row">
137
+ <label for="k">Optimizer Parameters (k):</label>
138
+ <input type="range" id="k" name="k" min="1" max="16" value="8">
139
+ <input type="number" id="k_input" value="8" min="1" max="16">
140
+ </div>
141
+ <div class="row">
142
+ <label for="tp">Tensor Model Parallelism (t):</label>
143
+ <input type="range" id="tp" name="tp" min="1" max="16" value="8">
144
+ <input type="number" id="tp_input" value="8" min="1" max="16">
145
+ </div>
146
+ <div class="row">
147
+ <label for="dp">Data Model Parallelism (d):</label>
148
+ <input type="range" id="dp" name="dp" min="1" max="256" value="1">
149
+ <input type="number" id="dp_input" value="1" min="1" max="256">
150
+ </div>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
151
  </div>
152
+ <div class="column">
153
+ <div class="row">
154
+ <label for="mixed">Mixed Precision:</label>
155
+ <input type="checkbox" id="mixed" name="mixed" checked>
156
+ <span></span> <!-- Empty span to maintain grid alignment -->
157
+ </div>
158
+ <div class="row">
159
+ <label for="recomputation">Recomputation:</label>
160
+ <select id="recomputation" name="recomputation">
161
+ <option value="none">None</option>
162
+ <option value="selective">Selective</option>
163
+ <option value="full">Full</option>
164
+ </select>
165
+ <span></span> <!-- Empty span to maintain grid alignment -->
166
+ </div>
167
+ <div class="row">
168
+ <label for="zero">Zero:</label>
169
+ <select id="zero" name="zero">
170
+ <option value="0">0</option>
171
+ <option value="1">1</option>
172
+ <option value="2">2</option>
173
+ <option value="3">3</option>
174
+ </select>
175
+ <span></span> <!-- Empty span to maintain grid alignment -->
176
+ </div>
177
+ <div class="row">
178
+ <label for="ff_activation">FF Activation:</label>
179
+ <select id="ff_activation" name="ff_activation">
180
+ <option value="relu">ReLU</option>
181
+ <option value="gelu">GELU</option>
182
+ <option value="swiglu">SwiGLU</option>
183
+ </select>
184
+ <span></span> <!-- Empty span to maintain grid alignment -->
185
+ </div>
186
+ <div class="row">
187
+ <label for="presets">Presets:</label>
188
+ <select id="presets" name="presets">
189
+ <option value="Llama 3 Tiny">Llama 3 Tiny</option>
190
+ <option value="Llama 3 8B">Llama 3 8B</option>
191
+ <option value="Llama 3 70B">Llama 3 70B</option>
192
+ <option value="Llama 3 405B">Llama 3 405B</option>
193
+ </select>
194
+ <span></span> <!-- Empty span to maintain grid alignment -->
195
+ </div>
196
  </div>
197
  </div>
198
 
dist/main.bundle.js CHANGED
The diff for this file is too large to render. See raw diff
 
dist/main.bundle.js.map CHANGED
The diff for this file is too large to render. See raw diff
 
src/index.html CHANGED
@@ -9,32 +9,39 @@
9
  <title>FineWeb: decanting the web for the finest text data at scale</title>
10
  <link rel="stylesheet" href="style.css">
11
  <style>
12
- #controls {
13
- display: grid;
14
- grid-template-columns: auto 1fr auto;
15
- gap: 5px;
16
- align-items: center;
17
- max-width: 600px;
18
- margin-bottom: 20px;
19
- }
20
- #controls label {
21
- text-align: right;
22
- }
23
- #controls input[type="range"] {
24
- width: 100%;
25
- }
26
- #controls input[type="number"] {
27
- width: 60px;
28
- }
29
- #controls .row {
30
- display: contents;
31
- }
32
- #graph svg {
33
- font-family: sans-serif;
34
- }
35
- #graph svg rect {
36
- cursor: pointer;
37
- }
 
 
 
 
 
 
 
38
  </style>
39
  </head>
40
 
@@ -90,98 +97,102 @@
90
 
91
  <div id="graph" style="position: relative; width: 700px; height: 500px;"></div>
92
  <div id="controls">
93
- <div class="row">
94
- <label for="a">Attention Heads (a):</label>
95
- <input type="range" id="a" name="a" min="1" max="128" value="8">
96
- <input type="number" id="a_input" value="8" min="1" max="128">
97
- </div>
98
- <div class="row">
99
- <label for="b">Micro Batch Size (b):</label>
100
- <input type="range" id="b" name="b" min="1" max="53248" value="32">
101
- <input type="number" id="b_input" value="32" min="1" max="53248">
102
- </div>
103
- <div class="row">
104
- <label for="h">Hidden Dimension Size (h):</label>
105
- <input type="range" id="h" name="h" min="1" max="16384" value="512">
106
- <input type="number" id="h_input" value="512" min="128" max="16384">
107
- </div>
108
- <div class="row">
109
- <label for="h_ff">Feedforward Dimension Size (h_ff):</label>
110
- <input type="range" id="h_ff" name="h_ff" min="1" max="65536" value="2048">
111
- <input type="number" id="h_ff_input" value="2048" min="512" max="65536">
112
- </div>
113
- <div class="row">
114
- <label for="L">Number of Layers (L):</label>
115
- <input type="range" id="L" name="L" min="1" max="126" value="12">
116
- <input type="number" id="L_input" value="12" min="1" max="126">
117
- </div>
118
- <div class="row">
119
- <label for="s">Sequence Length (s):</label>
120
- <input type="range" id="s" name="s" min="1" max="128000" value="128">
121
- <input type="number" id="s_input" value="128" min="64" max="128000">
122
- </div>
123
- <div class="row">
124
- <label for="v">Vocabulary Size (v):</label>
125
- <input type="range" id="v" name="v" min="1000" max="100000" value="30522">
126
- <input type="number" id="v_input" value="30522" min="1000" max="100000">
127
- </div>
128
- <div class="row">
129
- <label for="k">Optimizer Parameters (k):</label>
130
- <input type="range" id="k" name="k" min="1" max="16" value="8">
131
- <input type="number" id="k_input" value="8" min="1" max="16">
132
- </div>
133
- <div class="row">
134
- <label for="tp">Tensor Model Parallelism (t):</label>
135
- <input type="range" id="tp" name="tp" min="1" max="16" value="8">
136
- <input type="number" id="tp_input" value="8" min="1" max="16">
137
- </div>
138
- <div class="row">
139
- <label for="dp">Data Model Parallelism (d):</label>
140
- <input type="range" id="dp" name="dp" min="1" max="256" value="1">
141
- <input type="number" id="dp_input" value="1" min="1" max="256">
142
- </div>
143
- <div class="row">
144
- <label for="mixed">Mixed Precision:</label>
145
- <input type="checkbox" id="mixed" name="mixed" checked>
146
- <span></span> <!-- Empty span to maintain grid alignment -->
147
- </div>
148
- <div class="row">
149
- <label for="recomputation">Recomputation:</label>
150
- <select id="recomputation" name="recomputation">
151
- <option value="none">None</option>
152
- <option value="selective">Selective</option>
153
- <option value="full">Full</option>
154
- </select>
155
- <span></span> <!-- Empty span to maintain grid alignment -->
156
- </div>
157
- <div class="row">
158
- <label for="zero">Zero:</label>
159
- <select id="zero" name="zero">
160
- <option value="0">0</option>
161
- <option value="1">1</option>
162
- <option value="2">2</option>
163
- <option value="3">3</option>
164
- </select>
165
- <span></span> <!-- Empty span to maintain grid alignment -->
166
- </div>
167
- <div class="row">
168
- <label for="ff_activation">FF Activation:</label>
169
- <select id="ff_activation" name="ff_activation">
170
- <option value="relu">ReLU</option>
171
- <option value="gelu">GELU</option>
172
- <option value="swiglu">SwiGLU</option>
173
- </select>
174
- <span></span> <!-- Empty span to maintain grid alignment -->
175
  </div>
176
- <div class="row">
177
- <label for="presets">Presets:</label>
178
- <select id="presets" name="presets">
179
- <option value="Llama 3 Tiny">Llama 3 Tiny</option>
180
- <option value="Llama 3 8B">Llama 3 8B</option>
181
- <option value="Llama 3 70B">Llama 3 70B</option>
182
- <option value="Llama 3 405B">Llama 3 405B</option>
183
- </select>
184
- <span></span> <!-- Empty span to maintain grid alignment -->
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
185
  </div>
186
  </div>
187
 
 
9
  <title>FineWeb: decanting the web for the finest text data at scale</title>
10
  <link rel="stylesheet" href="style.css">
11
  <style>
12
+ #controls {
13
+ display: grid;
14
+ grid-template-columns: auto 1fr auto;
15
+ gap: 5px;
16
+ align-items: center;
17
+ max-width: 700px;
18
+ margin-bottom: 10px;
19
+ }
20
+ #controls .row {
21
+ display: contents;
22
+ }
23
+ #controls label {
24
+ text-align: right;
25
+ padding-right: 10px;
26
+ }
27
+ #controls input[type="range"] {
28
+ width: 100%;
29
+ }
30
+ #controls input[type="number"] {
31
+ width: 60px;
32
+ }
33
+ #controls select {
34
+ width: 100%;
35
+ }
36
+ #controls .column {
37
+ display: contents;
38
+ }
39
+ #graph svg {
40
+ font-family: sans-serif;
41
+ }
42
+ #graph svg rect {
43
+ cursor: pointer;
44
+ }
45
  </style>
46
  </head>
47
 
 
97
 
98
  <div id="graph" style="position: relative; width: 700px; height: 500px;"></div>
99
  <div id="controls">
100
+ <div class="column">
101
+ <div class="row">
102
+ <label for="a">Attention Heads (a):</label>
103
+ <input type="range" id="a" name="a" min="1" max="128" value="8">
104
+ <input type="number" id="a_input" value="8" min="1" max="128">
105
+ </div>
106
+ <div class="row">
107
+ <label for="b">Micro Batch Size (b):</label>
108
+ <input type="range" id="b" name="b" min="1" max="53248" value="32">
109
+ <input type="number" id="b_input" value="32" min="1" max="53248">
110
+ </div>
111
+ <div class="row">
112
+ <label for="h">Hidden Dimension Size (h):</label>
113
+ <input type="range" id="h" name="h" min="1" max="16384" value="512">
114
+ <input type="number" id="h_input" value="512" min="128" max="16384">
115
+ </div>
116
+ <div class="row">
117
+ <label for="h_ff">Feedforward Dimension Size (h_ff):</label>
118
+ <input type="range" id="h_ff" name="h_ff" min="1" max="65536" value="2048">
119
+ <input type="number" id="h_ff_input" value="2048" min="512" max="65536">
120
+ </div>
121
+ <div class="row">
122
+ <label for="L">Number of Layers (L):</label>
123
+ <input type="range" id="L" name="L" min="1" max="126" value="12">
124
+ <input type="number" id="L_input" value="12" min="1" max="126">
125
+ </div>
126
+ <div class="row">
127
+ <label for="s">Sequence Length (s):</label>
128
+ <input type="range" id="s" name="s" min="1" max="128000" value="128">
129
+ <input type="number" id="s_input" value="128" min="64" max="128000">
130
+ </div>
131
+ <div class="row">
132
+ <label for="v">Vocabulary Size (v):</label>
133
+ <input type="range" id="v" name="v" min="1000" max="100000" value="30522">
134
+ <input type="number" id="v_input" value="30522" min="1000" max="100000">
135
+ </div>
136
+ <div class="row">
137
+ <label for="k">Optimizer Parameters (k):</label>
138
+ <input type="range" id="k" name="k" min="1" max="16" value="8">
139
+ <input type="number" id="k_input" value="8" min="1" max="16">
140
+ </div>
141
+ <div class="row">
142
+ <label for="tp">Tensor Model Parallelism (t):</label>
143
+ <input type="range" id="tp" name="tp" min="1" max="16" value="8">
144
+ <input type="number" id="tp_input" value="8" min="1" max="16">
145
+ </div>
146
+ <div class="row">
147
+ <label for="dp">Data Model Parallelism (d):</label>
148
+ <input type="range" id="dp" name="dp" min="1" max="256" value="1">
149
+ <input type="number" id="dp_input" value="1" min="1" max="256">
150
+ </div>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
151
  </div>
152
+ <div class="column">
153
+ <div class="row">
154
+ <label for="mixed">Mixed Precision:</label>
155
+ <input type="checkbox" id="mixed" name="mixed" checked>
156
+ <span></span> <!-- Empty span to maintain grid alignment -->
157
+ </div>
158
+ <div class="row">
159
+ <label for="recomputation">Recomputation:</label>
160
+ <select id="recomputation" name="recomputation">
161
+ <option value="none">None</option>
162
+ <option value="selective">Selective</option>
163
+ <option value="full">Full</option>
164
+ </select>
165
+ <span></span> <!-- Empty span to maintain grid alignment -->
166
+ </div>
167
+ <div class="row">
168
+ <label for="zero">Zero:</label>
169
+ <select id="zero" name="zero">
170
+ <option value="0">0</option>
171
+ <option value="1">1</option>
172
+ <option value="2">2</option>
173
+ <option value="3">3</option>
174
+ </select>
175
+ <span></span> <!-- Empty span to maintain grid alignment -->
176
+ </div>
177
+ <div class="row">
178
+ <label for="ff_activation">FF Activation:</label>
179
+ <select id="ff_activation" name="ff_activation">
180
+ <option value="relu">ReLU</option>
181
+ <option value="gelu">GELU</option>
182
+ <option value="swiglu">SwiGLU</option>
183
+ </select>
184
+ <span></span> <!-- Empty span to maintain grid alignment -->
185
+ </div>
186
+ <div class="row">
187
+ <label for="presets">Presets:</label>
188
+ <select id="presets" name="presets">
189
+ <option value="Llama 3 Tiny">Llama 3 Tiny</option>
190
+ <option value="Llama 3 8B">Llama 3 8B</option>
191
+ <option value="Llama 3 70B">Llama 3 70B</option>
192
+ <option value="Llama 3 405B">Llama 3 405B</option>
193
+ </select>
194
+ <span></span> <!-- Empty span to maintain grid alignment -->
195
+ </div>
196
  </div>
197
  </div>
198
 
src/memory.js CHANGED
@@ -294,10 +294,10 @@ function formatBytes(bytes) {
294
  }
295
 
296
  const presets = {
297
- "Llama 3 Tiny": { a: 16, b: 3, h: 1024, h_ff: 4096, L: 1, s: 7, v: 30522, k: 8, tp: 1, zero: "Optimizer", dp: 1, mixed: true, recomputation: "none", ff_activation: "gelu" },
298
- "Llama 3 8B": { a: 32, b: 32, h: 4096, h_ff: 16384, L: 32, s: 256, v: 30522, k: 8, tp: 1, zero: "Optimizer", dp: 1, mixed: true, recomputation: "none", ff_activation: "swiglu" },
299
- "Llama 3 70B": { a: 64, b: 32, h: 8192, h_ff: 32768, L: 80, s: 256, v: 30522, k: 8, tp: 1, zero: "Optimizer", dp: 1, mixed: true, recomputation: "none", ff_activation: "swiglu" },
300
- "Llama 3 405B": { a: 128, b: 32, h: 16384, h_ff: 65536, L: 126, s: 256, v: 30522, k: 8, t: 1, mixed: true, recomputation: "none", ff_activation: "swiglu" }
301
  };
302
 
303
  function setPresetValues(preset) {
 
294
  }
295
 
296
  const presets = {
297
+ "Llama 3 Tiny": { a: 16, b: 3, h: 1024, h_ff: 4096, L: 1, s: 7, v: 30522, k: 8, tp: 1, zero: "1", dp: 1, mixed: true, recomputation: "none", ff_activation: "gelu" },
298
+ "Llama 3 8B": { a: 32, b: 32, h: 4096, h_ff: 16384, L: 32, s: 256, v: 30522, k: 8, tp: 1, zero: "1", dp: 1, mixed: true, recomputation: "none", ff_activation: "swiglu" },
299
+ "Llama 3 70B": { a: 64, b: 32, h: 8192, h_ff: 32768, L: 80, s: 256, v: 30522, k: 8, tp: 8, zero: "1", dp: 8, mixed: true, recomputation: "none", ff_activation: "swiglu" },
300
+ "Llama 3 405B": { a: 128, b: 32, h: 16384, h_ff: 65536, L: 126, s: 256, v: 30522, k: 8, tp: 8, zero: "1", dp: 8, mixed: true, recomputation: "none", ff_activation: "swiglu" }
301
  };
302
 
303
  function setPresetValues(preset) {