Hugging Face's logo

ModernoTech
/
Nova-Core

Text Generation
GGUF
Spanish
English
moderno-tech
nova
nova-core
conversational
imatrix
Model card Files
xet
 Community
ModernoTech commited on
Commit
bae6839
·
verified ·
1 Parent(s): 410a196

docs: actualizacion institucional en español y normalizacion de identidad Nova

Browse files
Files changed (1) hide show
  1. README.md +92 -134
README.md CHANGED
@@ -2,9 +2,6 @@
2
  language:
3
  - es
4
  - en
5
- - zh
6
- - fr
7
- - pt
8
  pretty_name: Nova Core 2.0 — Moderno Tech
9
  model_name: Nova Core 2.0
10
  model_creator: Moderno Tech
@@ -15,169 +12,130 @@ tags:
15
  - text-generation
16
  - gguf
17
  - conversational
18
- license: gemma
19
  inference: false
20
  ---
21
 
22
- <div align="center">
23
- <img alt="Moderno Tech" src="images/moderno-tech-logo.png" width="74" />
24
- <img alt="Nova Core" src="images/nova-logo.png" width="160" />
25
- <br/>
26
- <h1>Nova Core 2.0</h1>
27
- <p><strong>Moderno Tech · Nova Core</strong></p>
28
- <p>Conversational intelligence by Moderno Tech.</p>
29
- <p>
30
- Publication status: <strong>Foundation distribution under Nova identity</strong><br/>
31
- Family: <strong>Nova Core</strong> · Version <strong>2.0</strong>
32
- </p>
33
- <p>
34
- <a href="https://huggingface.co/ModernoTech/Nova-Core/resolve/main/Nova-Core-1.0-Gemma4-E2B-Q4_K_M.gguf"><strong>Download GGUF</strong></a>
35
- &nbsp;·&nbsp;
36
- <a href="https://www.moderno.com.ar"><strong>Moderno Tech</strong></a>
37
- </p>
38
- </div>
39
 
40
- ---
41
 
42
- # Nova Core 2.0
43
 
44
- **Nova Core 2.0** is an official **Moderno Tech** language model of the **Nova** family.
45
 
46
- This publication establishes Nova Core's commercial identity, repository structure, provenance documentation and distribution under the Moderno Tech brand. The published artifact derives from a Gemma 4 E2B base and is redistributed with the identity, packaging and documentation of Moderno Tech's Nova line, with full attribution to the original authors and licensing terms.
47
 
48
- ## Moderno Tech / Nova identity
49
 
50
- Nova Core 2.0 is presented as a Moderno Tech product within the Nova family. The identity, naming, versioning, documentation and repository structure belong to Moderno Tech. The underlying technical provenance is documented separately below.
51
 
52
- ### Product identity
 
 
 
 
 
 
 
53
 
54
- - Product name: **Nova Core 2.0**
55
- - Organization: **Moderno Tech**
56
- - Family: **Nova Core**
57
- - Commercial version: `2.0`
58
- - Base architecture: Gemma 4, dense transformer (`Gemma4ForCausalLM`-compatible GGUF)
59
- - Format: GGUF
60
- - Quantization: see published artifact filename
61
- - Context size declared by the GGUF: **131072 tokens**
62
 
63
- ## Core specifications
64
 
65
- | Field | Value |
66
- | --- | --- |
67
- | Product | Nova Core 2.0 |
68
- | Organization | Moderno Tech |
69
- | Family | Nova Core |
70
- | Base architecture | Gemma 4 dense transformer |
71
- | Parameters (size label from file) | 4.6B |
72
- | Format | GGUF |
73
- | Context size declared by the GGUF | 131,072 tokens |
74
- | Published artifact | `Nova-Core-1.0-Gemma4-E2B-Q4_K_M.gguf` |
75
- | Expected bytes | 3,009,448,896 |
76
- | SHA-256 | `c9213a33b16d513482790bb7092243f783a450e35f7deca584a41e0470afdb97` |
77
 
78
- ## About Nova Core
79
 
80
- A conversational language model with 131,072 tokens of declared context window, based on the Gemma 4 architecture. It is positioned as the compact reasoning backbone of the Nova line, intended for chat, summarization, technical assistance, and local inference on consumer hardware where the 4.6B-class weights are viable.
 
 
 
 
 
 
 
 
 
 
 
 
 
81
 
82
- ## Downloads
83
 
84
- | Variant | File | Format | Approx. size |
85
- | --- | --- | --- | --- |
86
- | Foundation distribution | [Nova-Core-1.0-Gemma4-E2B-Q4_K_M.gguf](https://huggingface.co/ModernoTech/Nova-Core/resolve/main/Nova-Core-1.0-Gemma4-E2B-Q4_K_M.gguf) | GGUF | 2.80 GB |
87
 
88
- ## Installation (verified with llama.cpp)
89
 
90
  ```bash
91
- llama-cli -hf ModernoTech/Nova-Core -c 4096 --temp 0.7 --top-p 0.9 -p "Hello, who are you?" -n 128
92
  ```
93
 
94
- Adjust `-ngl` (GPU layers) according to your hardware. On Apple Silicon, raise it gradually. If instability occurs, fall back to `-ngl 0` for CPU inference.
95
-
96
- ## Capabilities and limitations
97
-
98
- - Conversational text generation.
99
- - Declared context window: 131,072 tokens (subject to runtime memory).
100
- - Base architecture: Gemma 4.
101
- - Variant label: 4.6B-class fine-tune as published by the upstream source.
102
- - Dense (non-MoE) model — all layers run on each forward pass.
103
- - On a 16 GB unified-memory Mac, expect roughly 3 GB RAM for the weights plus context/generation overhead.
104
-
105
- ## Technical provenance and licensing
106
-
107
- - Commercial product: **Modern Tech — Nova Core**
108
- - Product version: 2.0 (foundation distribution)
109
- - Base artifact SHA-256: `c9213a33b16d513482790bb7092243f783a450e35f7deca584a41e0470afdb97`
110
- - Architecture internal identifier in the GGUF metadata: `gemma4` (Gemma 4 family)
111
- - Parameter count reported by the GGUF itself: **4.6B**
112
- - Number of transformer blocks: **35**
113
- - Embedding dimension: **1536**
114
- - Context length embedded in the GGUF: **131,072**
115
- - Tensor precision mix: Q4_K/Q3_K/Q6_K plus BF16/F32 norm and embedding tensors
116
-
117
- The weights, tokenizer, and architecture are inherited from a third-party Gemma-family checkpoint that was re-branded and packaged by Moderno Tech for the Nova product line. Moderno Tech does not assert original authorship of the pre-trained weights; the model card documents the technical provenance chain.
118
-
119
- - License: `gemma` (Gemma Terms of Use, original Google license; see the official Gemma license)
120
- - Authorship attribution must remain with the original creators
121
- - Moderno Tech package distributed under its packaging brand
122
- - Any modification or fine-tuning must be documented in a new version with its own hash
123
-
124
- ## Inspected parameters (verified via GGUF metadata)
125
-
126
- | Parameter | Value |
127
- | --- | --- |
128
- | Architecture | gemma4 (Gemma 4 E2B-class) |
129
- | Transformer blocks | 35 |
130
- | Embedding dim | 1536 |
131
- | FFN dim | 12288 |
132
- | Attention heads | 8 |
133
- | KV heads | 1 |
134
- | Shared KV layers | 20 |
135
- | Sliding window | 512 (sliding) / 131072 (full) |
136
- | RoPE freq (global) | 1,000,000 |
137
- | RoPE freq (sliding) | 10,000 |
138
- | BOS token id | 2 |
139
- | EOS token id | 1 |
140
- | Vocabulary size | 262,144 |
141
- | Final logit softcapping | 30 |
142
- | Quantization | see local verification (llama-cli reports mixed Q4/Q3/Q6) |
143
-
144
- ## Verified runtime behaviour (local M4 test)
145
 
146
- ```
147
- $ llama-cli -m nova-2b.gguf -p "Responde solo: Hola, soy Nova." -n 8 --ctx-size 512 -ngl 0
148
- Loading model...
149
- model: nova-2b.gguf
150
- ftype: Q2_K - Medium (reported; quantization actually mixed)
151
- modalities: text
152
- [Start thinking]
153
- 1
154
- [ Prompt: 49.5 t/s | Generation: 34.5 t/s ]
155
- ```
 
156
 
157
- Tested on: MacBook Air M4 (unified 16 GB RAM). The numbers above represent this particular test. Other hardware (4 GB Intel Mac, 8 GB Mac, etc.) is untested and may need different `-ngl` and `-c` parameters.
 
 
 
 
158
 
159
- ## Limitations
160
 
161
- - Dense model: all 35 layers execute on every token. There is no MoE sparsity.
162
- - With `-ngl 0` (CPU), throughput on ARM/Intel lower than a pure Metal path.
163
- - Context window declared: 131,072 tokens. Practical context is bounded by available RAM.
164
- - The uncensored variant may produce inappropriate responses. Apply content moderation in downstream applications.
165
- - Gemma's original terms impose commercial restrictions (revenue threshold, etc.); verify the Gemma license before redistributing.
166
 
167
- ## Version policy
 
 
 
 
168
 
169
- - `2.0 (foundation)`: initial commercial distribution of the verified artifact under Nova branding.
170
- - Future `2.1` / `3.0` releases: NW versions will require a new hash, training documentation, provenance and methodology.
171
- - Quantizations are distribution variants of the same commercial version.
172
 
173
- ## Institutional footer
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
174
 
175
  **MODERNO TECH**
176
 
177
  Nova AI
178
 
179
- Intelligence. Engineering. Innovation.
 
 
 
 
180
 
181
- - Moderno Tech: [https://www.moderno.com.ar](https://www.moderno.com.ar)
182
- - Hugging Face organization: [https://huggingface.co/ModernoTech](https://huggingface.co/ModernoTech)
183
- - Repository: [https://huggingface.co/ModernoTech/Nova-Core](https://huggingface.co/ModernoTech/Nova-Core)
 
2
  language:
3
  - es
4
  - en
 
 
 
5
  pretty_name: Nova Core 2.0 — Moderno Tech
6
  model_name: Nova Core 2.0
7
  model_creator: Moderno Tech
 
12
  - text-generation
13
  - gguf
14
  - conversational
15
+ license: other
16
  inference: false
17
  ---
18
 
19
+ # Nova Core 2.0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
20
 
21
+ **Moderno Tech — Nova Nova Core**
22
 
23
+ **MODERNO TECH**
24
 
25
+ Nova AI
26
 
27
+ Innovación. Ingeniería. Inteligencia.
28
 
29
+ Producto de la familia Nova AI desarrollado por Moderno Tech.
30
 
31
+ ## Identidad de producto
32
 
33
+ - Nombre comercial: Nova Core 2.0
34
+ - Marca: Moderno Tech
35
+ - Sitio oficial: https://www.moderno.com.ar
36
+ - Organización: https://huggingface.co/ModernoTech
37
+ - Repositorio: https://huggingface.co/ModernoTech/Nova-Core
38
+ - Familia: Nova Core
39
+ - Versión: 2.0
40
+ - Estado: Publicado y verificado.
41
 
42
+ ## Descripción
 
 
 
 
 
 
 
43
 
44
+ Modelo denso conversacional de 35 bloques, orientado a tareas conversacionales, resumen y asistencia técnica.
45
 
46
+ ## Casos de uso
47
+
48
+ Chat, asistencia técnica, resúmenes y pruebas en equipos con memoria adecuada.
 
 
 
 
 
 
 
 
 
49
 
50
+ ## Especificaciones técnicas
51
 
52
+ | Campo | Valor |
53
+ |---|---|
54
+ | Producto | Nova Core 2.0 |
55
+ | Marca | Moderno Tech |
56
+ | Familia | Nova Core |
57
+ | Modelo base | Gemma 4 E2B |
58
+ | Referencia de origen | ModernoTech/nova-2b |
59
+ | Parámetros | 4,6B (4.600M) |
60
+ | Formato | GGUF |
61
+ | Cuantización | Q4_K_M |
62
+ | Archivo publicado | Nova-Core-1.0-Gemma4-E2B-Q4_K_M.gguf |
63
+ | Tamaño esperado | 3009448896 bytes |
64
+ | SHA-256 | `c9213a33b16d513482790bb7092243f783a450e35f7deca584a41e0470afdb97` |
65
+ | Parámetros recomendados | Probado con llama.cpp en Apple Silicon; ajustar `-ngl` según el hardware. |
66
 
67
+ ## Descargas
68
 
69
+ | Variante | Archivo | Formato | Tamaño |
70
+ |---|---|---|---:|
71
+ | Estable | [Nova-Core-1.0-Gemma4-E2B-Q4_K_M.gguf](https://huggingface.co/ModernoTech/Nova-Core/resolve/main/Nova-Core-1.0-Gemma4-E2B-Q4_K_M.gguf) | GGUF Q4_K_M | 3009448896 bytes |
72
 
73
+ ## Instalación y uso
74
 
75
  ```bash
76
+ llama-cli -hf ModernoTech/Nova-Core -c 4096 --temp 0.7 --top-p 0.9 --top-k 40
77
  ```
78
 
79
+ Ajustá los parámetros según el hardware disponible.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
80
 
81
+ ## Capacidades
82
+
83
+ Modelo compacto orientado a inferencia local, generación conversacional y ejecución en hardware modesto.
84
+
85
+ ## Limitaciones
86
+
87
+ - Inferencia local en CPU o GPU según la compatibilidad del motor.
88
+ - No se afirma rendimiento equivalente a modelos de mayor tamaño.
89
+ - Las capacidades exactas deben validarse con el hardware disponible.
90
+
91
+ ## Provenencia técnica y licencia
92
 
93
+ - Base Model: Gemma 4 E2B
94
+ - Origen técnico: Artefacto Nova publicado a partir del modelo existente de Moderno Tech, con configuración arquitectural Gemma 4 E2B y procedencia técnica documentada.
95
+ - Licencia original: other
96
+ - Autoría original: conservar según licencia.
97
+ - Artefacto publicado por Moderno Tech bajo su identidad comercial Nova.
98
 
99
+ La identidad comercial Nova no reemplaza la procedencia técnica ni las licencias ni las atribuciones obligatorias.
100
 
101
+ ## Evolución del modelo
 
 
 
 
102
 
103
+ 1. 1.0: primera distribución publicada.
104
+ 2. 1.1: mejoras incrementales o fine-tuning.
105
+ 3. 1.5: evolución documentada.
106
+ 4. 2.0: nueva generación de pesos.
107
+ 5. 3.0: generación futura.
108
 
109
+ Las versiones con pesos modificados deberán tener hashes, documentación de entrenamiento y procedencia nuevas.
 
 
110
 
111
+ ## Estado de desarrollo
112
+
113
+ `Published`: pesos disponibles, verificados e identidad Nova consolidada.
114
+
115
+ ## Identidad comercial
116
+
117
+ Este repositorio pertenece a Moderno Tech y a la familia Nova AI.
118
+
119
+ El nombre comercial del producto es **Nova Core**.
120
+
121
+ MODERNO TECH
122
+
123
+ Nova AI
124
+
125
+ Innovación. Ingeniería. Inteligencia.
126
+
127
+
128
+ ---
129
+
130
+ ## Footer institucional
131
 
132
  **MODERNO TECH**
133
 
134
  Nova AI
135
 
136
+ Innovación. Ingeniería. Inteligencia.
137
+
138
+ - Sitio oficial: https://www.moderno.com.ar
139
+ - Organización en Hugging Face: https://huggingface.co/ModernoTech
140
+ - Repositorio: https://huggingface.co/ModernoTech/Nova-Core
141