File size: 6,540 Bytes
35398ce 82db140 35398ce 1c5bb76 35398ce 1c5bb76 35398ce 1c5bb76 35398ce 4f81d6b 35398ce 1c5bb76 35398ce 82db140 35398ce dda4ae5 35398ce cd76cee 35398ce |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 |
---
license: apache-2.0
datasets:
- GreenerPastures/All-Your-Base-Full
language:
- en
base_model:
- Qwen/Qwen3-4B-Base
---
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>Basically-Human-4B</title>
<link href="https://fonts.googleapis.com/css2?family=Share+Tech+Mono&display=swap" rel="stylesheet">
<style>
body {
font-family: 'Share Tech Mono', monospace;
background-color: #0f0f1a;
color: #ccc;
margin: 0;
padding: 0;
}
.container {
width: 95%;
margin: 20px auto;
background-color: #1a1a2e;
padding: 30px;
border-radius: 12px;
box-shadow: 0 0 20px #00fff7;
}
h1, h2, h3, h4 {
font-family: 'Share Tech Mono', monospace;
color: #00fff7;
text-align: center;
}
h3 {
font-size: 3em;
margin: 20px 0;
text-shadow: 0 0 10px #00fff7;
}
p {
line-height: 1.7;
font-size: 1.1em;
color: #ccc;
}
.image-container {
margin-bottom: 30px;
}
.image-container img {
display: block;
margin: 0 auto;
border-radius: 8px;
box-shadow: 0 0 15px #00fff7;
max-width: 100%;
height: auto;
}
.description, .procedure, .data, .quantizations {
margin-bottom: 50px;
}
.list {
list-style-type: square;
margin-left: 25px;
color: #0ff;
}
.quote {
background-color: #292945;
padding: 15px;
border-left: 4px solid #00fff7;
font-style: italic;
margin: 20px 0;
color: #eee;
}
.section-title {
font-size: 1.8em;
color: #00fff7;
border-bottom: 2px solid #00fff7;
padding-bottom: 8px;
margin-bottom: 15px;
}
.quantization-info {
background-color: #222237;
padding: 20px;
border-radius: 10px;
box-shadow: 0 0 15px #0ff;
}
.quantization-info h4 {
font-family: 'Share Tech Mono', monospace;
color: #00fff7;
margin-bottom: 10px;
}
a {
color: #0ff;
text-decoration: none;
}
a:hover {
color: #ff00cc;
text-decoration: underline;
}
</style>
</head>
<body>
<div class="container">
<h3>Basically-Human-4B</h3>
<div class="image-container">
<img src="https://cdn-uploads.huggingface.co/production/uploads/67db37734ef7ccf7781e67ad/4P0r7TL8xRsUM4b1huIDi.png" alt="Basically-Human-4B Image" width="500" />
</div>
<div class="description">
<h2 class="section-title">Description</h2>
<p>Basically-Human-4B is a 4B parameter roleplaying language model built on Qwen3 4B, fine-tuned for immersive, emotionally resonant character interaction. Designed to emulate the flow and nuance of real human dialogue, it excels at staying in persona, crafting believable responses, and driving dynamic storytelling. With support for ChatML formatting and alignment focused on realism and consistency, Basically-Human-4B is a compact but capable choice for text-based roleplay, NPC simulation, and interactive fiction.</p>
</div>
<div class="description">
<h2 class="section-title">Instruct Format</h2>
<p>Basically-Human-4B uses the ChatML instruction format and was trained under the Qwen3 chat template. Messages are structured with system, user, and assistant roles, each enclosed in <|im_start|> and <|im_end|> tokens, ensuring compatibility with applications that rely on structured conversational data. This format enables clear role delineation and seamless multi-turn interactions, making the model highly effective for in-character dialogue and roleplay scenarios.</p>
</div>
<div class="procedure">
<h2 class="section-title">Procedure</h2>
<p>The pretrained base model was fine-tuned on instruct and roleplaying data in a single step. Once the supervised fine-tuning was complete, a Kahneman-Tversky optimization was applied as a final alignment step.</p>
</div>
<div class="data">
<h2 class="section-title">Data</h2>
<p>The model was trained on a diverse collection of instruction and roleplaying data, including the following sets:</p>
<ul class="list">
<li>PocketDoc/Dans-MemoryCore-CoreCurriculum-Small</li>
<li>NewEden/Kalo-Opus-Instruct-22k-Refusal-Murdered</li>
<li>Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned</li>
<li>NewEden/Gryphe-Sonnet-3.5-35k-Subset</li>
<li>Nitral-AI/GU_Instruct-ShareGPT</li>
<li>Nitral-AI/Medical_Instruct-ShareGPT</li>
<li>AquaV/Resistance-Sharegpt</li>
<li>AquaV/US-Army-Survival-Sharegpt</li>
<li>Gryphe/Sonnet3.5-SlimOrcaDedupCleaned</li>
<li>Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned</li>
<li>ResplendentAI/bluemoon</li>
<li>hardlyworking/openerotica-freedomrp-sharegpt-system</li>
<li>MinervaAI/Aesir-Preview</li>
<li>anthracite-core/c2_logs_32k_v1.1</li>
<li>Nitral-AI/Creative_Writing-ShareGPT</li>
<li>PJMixers/lodrick-the-lafted_OpusStories-Story2Prompt-ShareGPT</li>
<li>NewEden/Opus-accepted-hermes-rejected-shuffled</li>
</ul>
</div>
<div class="quantizations">
<h2 class="section-title">Quantizations</h2>
<div class="quantization-info">
<h4>We provide GGUF quantized versions of Basically-Human-4B:</h4>
<ul>
<li><a href="https://huggingface.co/mradermacher/Basically-Human-4B-GGUF" target="_blank">Basically-Human-4B GGUF</a></li>
<li><a href="https://huggingface.co/mradermacher/Basically-Human-4B-i1-GGUF" target="_blank">Basically-Human-4B-i1 GGUF</a></li>
</ul>
</div>
</div>
</div>
</body>
</html> |