Skip to content

Commit 3deca02

Browse files
dusty-nvdusty-nv
authored
Merge pull request #288 from dusty-nv/dev
Dev
2 parents 7307fcd + d8adbd0 commit 3deca02

File tree

13 files changed

+614
-45
lines changed

13 files changed

+614
-45
lines changed

‎docs/portal/css/models.css‎

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -59,6 +59,13 @@
5959
background-position-y: 0px, -74px;
6060
}
6161

62+
.gemma-3-header {
63+
background-image: url('/portal/data/models/images/headers/gemma3.svg'), linear-gradient(90deg, hsl(218.8, 21.2%, 50.3%) 0%, rgb(66, 39, 141) 57%, rgba(63, 82, 166, 0.9) 100%);
64+
background-size: 48%, 100%;
65+
background-position-x: 520px, 0px;
66+
background-position-y: 30px, 0px;
67+
}
68+
6269
.hf-smol-header {
6370
background-image: linear-gradient(90deg, hsl(36.4, 69%, 53.1%) 0%, rgb(66, 39, 141) 56%, rgba(63, 82, 166, 0) 65%),
6471
url('/portal/data/models/images/headers/smol-v2.webp');

‎docs/portal/data/models/gemma.json‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
{
22
"gemma-2": {
3-
"name": "Google Gemma",
3+
"name": "Google Gemma 2",
44
"header": "gemma-header",
55
"max_context_len": {"placeholder": 8192},
66
"prefill_chunk": {"placeholder": 8192},

‎docs/portal/data/models/gemma3.json‎

Lines changed: 40 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,40 @@
1+
{
2+
"gemma-3": {
3+
"name": "Google Gemma 3",
4+
"header": "gemma-3-header",
5+
"max_context_len": {"placeholder": 4096},
6+
"prefill_chunk": {"placeholder": 4096},
7+
"blacklist": ["mlc", "llama_cpp"],
8+
"tags": ["llm"],
9+
"links": {
10+
"google": {
11+
"name": "Google",
12+
"url": "https://ai.google.dev/gemma"
13+
}
14+
}
15+
},
16+
"gemma-3-1b-it": {
17+
"name": "Gemma 3 1B",
18+
"url": "hf.co/google/gemma-3-1b-it",
19+
"blacklist": ["mlc", "llama_cpp"],
20+
"tags": ["gemma-3", "orin"]
21+
},
22+
"gemma-3-4b-it": {
23+
"name": "Gemma 3 4B",
24+
"url": "hf.co/google/gemma-3-4b-it",
25+
"blacklist": ["mlc", "llama_cpp"],
26+
"tags": ["gemma-3", "orin"]
27+
},
28+
"gemma-3-12b-it": {
29+
"name": "Gemma 3 12B",
30+
"url": "hf.co/google/gemma-3-12b-it",
31+
"blacklist": ["mlc", "llama_cpp"],
32+
"tags": ["gemma-3", "orin-nx", "agx-orin"]
33+
},
34+
"gemma-3-27b-it": {
35+
"name": "Gemma 3 27B",
36+
"url": "hf.co/google/gemma-3-27b-it",
37+
"blacklist": ["mlc", "llama_cpp"],
38+
"tags": ["gemma-3", "agx-orin"]
39+
}
40+
}

‎docs/portal/data/models/images/headers/gemma3.svg‎

Lines changed: 22 additions & 0 deletions
Loading

‎docs/portal/dist/db.json‎

Lines changed: 217 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -50,6 +50,29 @@
5050
"name": "AWQ TinyChat",
5151
"tags": "api"
5252
},
53+
"vllm": {
54+
"name": "vLLM",
55+
"tags": "api",
56+
"links": {
57+
"mlc": {
58+
"name": "vLLM",
59+
"url": "https://github.com/vllm-project/vllm"
60+
}
61+
}
62+
},
63+
"vllm:jp6": {
64+
"name": "dustynv/vllm:0.7.4-r36.4.0-cu128-24.04",
65+
"docker_image": "dustynv/vllm:0.7.4-r36.4.0-cu128-24.04",
66+
"docker_cmd": "vllm serve ${MODEL}",
67+
"docker_args": "--host=${SERVER_ADDR} --port=${SERVER_PORT} --dtype=auto --max-num-seqs=${MAX_BATCH_SIZE} --max-model-len=${MAX_CONTEXT_LEN} --gpu-memory-utilization=0.75",
68+
"docker_options": "-it --rm",
69+
"server_host": "0.0.0.0:9000",
70+
"tags": [
71+
"container",
72+
"vllm",
73+
"l4t-r36"
74+
]
75+
},
5376
"sudonim": {
5477
"docker_cmd": "sudonim serve",
5578
"docker_options": "-it --rm",
@@ -65,6 +88,13 @@
6588
],
6689
"help": "The inference API and type of quantization used."
6790
},
91+
"fp16": {
92+
"name": "fp16 (vLLM)",
93+
"tags": [
94+
"quantization",
95+
"vllm"
96+
]
97+
},
6898
"q4f16_ft": {
6999
"name": "q4f16_ft (MLC)",
70100
"tags": [
@@ -2648,6 +2678,192 @@
26482678
"max_context_len": 8192,
26492679
"prefill_chunk": 8192
26502680
},
2681+
"gemma-3": {
2682+
"name": "Google Gemma 3",
2683+
"header": "gemma-3-header",
2684+
"max_context_len": {
2685+
"placeholder": 4096
2686+
},
2687+
"prefill_chunk": {
2688+
"placeholder": 4096
2689+
},
2690+
"blacklist": [
2691+
"mlc",
2692+
"llama_cpp"
2693+
],
2694+
"tags": [
2695+
"llm"
2696+
],
2697+
"links": {
2698+
"google": {
2699+
"name": "Google",
2700+
"url": "https://ai.google.dev/gemma"
2701+
},
2702+
"hf": {
2703+
"name": "Hugging Face",
2704+
"color": "yellow",
2705+
"url": "Google-Gemma-3"
2706+
}
2707+
},
2708+
"url": "Google-Gemma-3"
2709+
},
2710+
"gemma-3-1b-it": {
2711+
"name": "Gemma 3 1B",
2712+
"url": "hf.co/google/gemma-3-1b-it",
2713+
"blacklist": [
2714+
"mlc",
2715+
"llama_cpp"
2716+
],
2717+
"tags": [
2718+
"gemma-3",
2719+
"orin"
2720+
],
2721+
"links": {
2722+
"hf": {
2723+
"name": "Hugging Face",
2724+
"color": "yellow",
2725+
"url": "hf.co/google/gemma-3-1b-it"
2726+
}
2727+
},
2728+
"created_at": "2025-03-10 12:09:00+00:00",
2729+
"last_modified": "2025-03-12 14:50:25+00:00"
2730+
},
2731+
"gemma-3-4b-it": {
2732+
"name": "Gemma 3 4B",
2733+
"url": "hf.co/google/gemma-3-4b-it",
2734+
"blacklist": [
2735+
"mlc",
2736+
"llama_cpp"
2737+
],
2738+
"tags": [
2739+
"gemma-3",
2740+
"orin"
2741+
],
2742+
"links": {
2743+
"hf": {
2744+
"name": "Hugging Face",
2745+
"color": "yellow",
2746+
"url": "hf.co/google/gemma-3-4b-it"
2747+
}
2748+
},
2749+
"created_at": "2025-02-20 21:20:07+00:00",
2750+
"last_modified": "2025-03-12 08:30:08+00:00"
2751+
},
2752+
"gemma-3-12b-it": {
2753+
"name": "Gemma 3 12B",
2754+
"url": "hf.co/google/gemma-3-12b-it",
2755+
"blacklist": [
2756+
"mlc",
2757+
"llama_cpp"
2758+
],
2759+
"tags": [
2760+
"gemma-3",
2761+
"orin-nx",
2762+
"agx-orin"
2763+
],
2764+
"links": {
2765+
"hf": {
2766+
"name": "Hugging Face",
2767+
"color": "yellow",
2768+
"url": "hf.co/google/gemma-3-12b-it"
2769+
}
2770+
},
2771+
"created_at": "2025-03-01 19:11:34+00:00",
2772+
"last_modified": "2025-03-12 08:30:33+00:00"
2773+
},
2774+
"gemma-3-27b-it": {
2775+
"name": "Gemma 3 27B",
2776+
"url": "hf.co/google/gemma-3-27b-it",
2777+
"blacklist": [
2778+
"mlc",
2779+
"llama_cpp"
2780+
],
2781+
"tags": [
2782+
"gemma-3",
2783+
"agx-orin"
2784+
],
2785+
"links": {
2786+
"hf": {
2787+
"name": "Hugging Face",
2788+
"color": "yellow",
2789+
"url": "hf.co/google/gemma-3-27b-it"
2790+
}
2791+
},
2792+
"created_at": "2025-03-01 19:10:19+00:00",
2793+
"last_modified": "2025-03-12 08:30:59+00:00"
2794+
},
2795+
"gemma-3-1b-it-fp16-vllm-jp6": {
2796+
"title": "Gemma 3 1B \u276f vLLM fp16 \u276f JetPack 6.1+",
2797+
"quantization": "fp16",
2798+
"tags": [
2799+
"gemma-3-1b-it",
2800+
"fp16",
2801+
"vllm:jp6"
2802+
]
2803+
},
2804+
"gemma-3-1b-it-q4_0-ollama-jp6": {
2805+
"title": "Gemma 3 1B \u276f ollama q4_0 \u276f JetPack 6.1+",
2806+
"quantization": "q4_0",
2807+
"tags": [
2808+
"gemma-3-1b-it",
2809+
"q4_0",
2810+
"ollama:jp6"
2811+
]
2812+
},
2813+
"gemma-3-4b-it-fp16-vllm-jp6": {
2814+
"title": "Gemma 3 4B \u276f vLLM fp16 \u276f JetPack 6.1+",
2815+
"quantization": "fp16",
2816+
"tags": [
2817+
"gemma-3-4b-it",
2818+
"fp16",
2819+
"vllm:jp6"
2820+
]
2821+
},
2822+
"gemma-3-4b-it-q4_0-ollama-jp6": {
2823+
"title": "Gemma 3 4B \u276f ollama q4_0 \u276f JetPack 6.1+",
2824+
"quantization": "q4_0",
2825+
"tags": [
2826+
"gemma-3-4b-it",
2827+
"q4_0",
2828+
"ollama:jp6"
2829+
]
2830+
},
2831+
"gemma-3-12b-it-fp16-vllm-jp6": {
2832+
"title": "Gemma 3 12B \u276f vLLM fp16 \u276f JetPack 6.1+",
2833+
"quantization": "fp16",
2834+
"tags": [
2835+
"gemma-3-12b-it",
2836+
"fp16",
2837+
"vllm:jp6"
2838+
]
2839+
},
2840+
"gemma-3-12b-it-q4_0-ollama-jp6": {
2841+
"title": "Gemma 3 12B \u276f ollama q4_0 \u276f JetPack 6.1+",
2842+
"quantization": "q4_0",
2843+
"tags": [
2844+
"gemma-3-12b-it",
2845+
"q4_0",
2846+
"ollama:jp6"
2847+
]
2848+
},
2849+
"gemma-3-27b-it-fp16-vllm-jp6": {
2850+
"title": "Gemma 3 27B \u276f vLLM fp16 \u276f JetPack 6.1+",
2851+
"quantization": "fp16",
2852+
"tags": [
2853+
"gemma-3-27b-it",
2854+
"fp16",
2855+
"vllm:jp6"
2856+
]
2857+
},
2858+
"gemma-3-27b-it-q4_0-ollama-jp6": {
2859+
"title": "Gemma 3 27B \u276f ollama q4_0 \u276f JetPack 6.1+",
2860+
"quantization": "q4_0",
2861+
"tags": [
2862+
"gemma-3-27b-it",
2863+
"q4_0",
2864+
"ollama:jp6"
2865+
]
2866+
},
26512867
"deepseek-r1-distill": {
26522868
"name": "DeepSeek R1",
26532869
"header": "deepseek-header",
@@ -4830,6 +5046,7 @@
48305046
"open_webui"
48315047
],
48325048
"child_order": [
5049+
"gemma-3",
48335050
"deepseek-r1-distill",
48345051
"qwen-2.5",
48355052
"llama-3",

‎docs/portal/dist/nanolab.css‎

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1519,6 +1519,13 @@ input[type="radio"].toggle {
15191519
background-position-y: 0px, -74px;
15201520
}
15211521

1522+
.gemma-3-header {
1523+
background-image: url('/portal/data/models/images/headers/gemma3.svg'), linear-gradient(90deg, hsl(218.8, 21.2%, 50.3%) 0%, rgb(66, 39, 141) 57%, rgba(63, 82, 166, 0.9) 100%);
1524+
background-size: 48%, 100%;
1525+
background-position-x: 520px, 0px;
1526+
background-position-y: 30px, 0px;
1527+
}
1528+
15221529
.hf-smol-header {
15231530
background-image: linear-gradient(90deg, hsl(36.4, 69%, 53.1%) 0%, rgb(66, 39, 141) 56%, rgba(63, 82, 166, 0) 65%),
15241531
url('/portal/data/models/images/headers/smol-v2.webp');

0 commit comments

Comments
 (0)