Upload folder using huggingface_hub

README.md

@@ -0,0 +1,66 @@
+
+---
+license: mit
+tags:
+- llama3
+- quantized
+- bits-8
+- dynamic-quantization
+- context-8000
+- layer-fusion-conceptual
+- tensor-fusion-conceptual
+- bias-removal
+- decode
+- coherence-enhancement
+- custom-code
+- grouping
+- reward-alignment
+- reasoning-tuned
+- safetensors
+---
+
+# xddd-processed
+
+Este repositorio incluye un modelo basado en `hghghgkskdmskdms/xddd` con las siguientes transformaciones aplicadas y características conceptuales documentadas por un script. El modelo se guarda en formato `safetensors`.
+- Cuantización dinámica a 8 bits.
+- **Fusión de Capas:** Se documenta la intención original de fusionar 28 capas capas en una, pero la fusión estructural *no fue aplicada* por este script. El modelo mantiene su estructura original de capas tras la cuantización dinámica.
+- **Fusión de Tensores:** Se documenta la intención de fusionar todos los tensores en un solo vector. El tamaño conceptual total es 394190218 elementos. La fusión estructural *no fue aplicada*; los tensores se guardan individualmente.
+- Eliminación de sesgos (puestos a cero).
+- Desactivación conceptual de censura.
+- Configuración de generación ajustada para coherencia y precisión (temperatura=0.7, top_p=0.9, repetition_penalty=1.2).
+- Definición conceptual de funciones de decodificación (tokens, parámetros, respuestas, layers, neuronas, tensores, arquitectura y un tensor fusionado conceptual).
+- max_position_embeddings: 8000.
+- Incluye configuraciones conceptuales para: Lógica de agrupación (tamaño=128), Alineación con mecanismos de recompensa, y Ajuste para mejorar el razonamiento.
+
+**Nota:** Este modelo ha sido cuantizado dinámicamente y tiene los sesgos puestos a cero. La fusión de capas y tensores *no fue aplicada estructuralmente*. Su compatibilidad puede variar. Las características conceptuales (agrupación, recompensa, razonamiento, funciones de decodificación) se reflejan en la configuración y README, pero su implementación activa durante la inferencia o entrenamiento depende del código de carga y uso posterior del modelo.
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+model = AutoModelForCausalLM.from_pretrained("jnjj/xddd-processed", trust_remote_code=True)
+tokenizer = AutoTokenizer.from_pretrained("jnjj/xddd-processed")
+
+messages = [
+    {"role": "system", "content": "Eres un asistente útil. Responde concisamente."},
+    {"role": "user", "content": "¿Qué es la cuantización en modelos de IA?"}
+]
+
+input_ids = tokenizer.apply_chat_template(
+    messages,
+    tokenize=True,
+    add_generation_prompt=True,
+    return_tensors="pt"
+)
+
+input_ids = input_ids.to(model.parameters().__next__().device)
+
+print("Generando respuesta...")
+output_ids = model.generate(
+    input_ids,
+    max_new_tokens=200,
+)
+
+response = tokenizer.decode(output_ids[0], skip_special_tokens=False)
+print("Respuesta:")
+print(response)
+
+```

config.json

@@ -0,0 +1,146 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bias_removal": true,
+  "bos_token_id": 128000,
+  "censorship": false,
+  "eos_token_id": [
+    128001,
+    128008,
+    128009
+  ],
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 8000,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pad_token_id": 128004,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.51.3",
+  "unsloth_version": "2025.2.15",
+  "use_cache": true,
+  "vocab_size": 128260,
+  "quantization": {
+    "method": "dynamic",
+    "bits": 8
+  },
+  "fusion": {
+    "layers_original": 28,
+    "details": "structural_fusion_not_applied_by_script"
+  },
+  "tensor_fusion": true,
+  "tensor_fusion_size": 394190218,
+  "generation_tuning": {
+    "max_length": 20,
+    "max_new_tokens": 100,
+    "min_length": 0,
+    "min_new_tokens": null,
+    "early_stopping": false,
+    "max_time": null,
+    "stop_strings": null,
+    "do_sample": true,
+    "num_beams": 1,
+    "num_beam_groups": 1,
+    "penalty_alpha": null,
+    "dola_layers": null,
+    "use_cache": true,
+    "cache_implementation": null,
+    "cache_config": null,
+    "return_legacy_cache": null,
+    "prefill_chunk_size": null,
+    "temperature": 0.7,
+    "top_k": 50,
+    "top_p": 0.9,
+    "min_p": null,
+    "typical_p": 1.0,
+    "epsilon_cutoff": 0.0,
+    "eta_cutoff": 0.0,
+    "diversity_penalty": 0.0,
+    "repetition_penalty": 1.2,
+    "encoder_repetition_penalty": 1.0,
+    "length_penalty": 1.0,
+    "no_repeat_ngram_size": 3,
+    "bad_words_ids": null,
+    "force_words_ids": null,
+    "renormalize_logits": false,
+    "constraints": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "remove_invalid_values": false,
+    "exponential_decay_length_penalty": null,
+    "suppress_tokens": null,
+    "begin_suppress_tokens": null,
+    "forced_decoder_ids": null,
+    "sequence_bias": null,
+    "token_healing": false,
+    "guidance_scale": null,
+    "low_memory": null,
+    "watermarking_config": null,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "output_logits": null,
+    "return_dict_in_generate": false,
+    "pad_token_id": null,
+    "bos_token_id": null,
+    "eos_token_id": null,
+    "encoder_no_repeat_ngram_size": 0,
+    "decoder_start_token_id": null,
+    "is_assistant": false,
+    "num_assistant_tokens": 20,
+    "num_assistant_tokens_schedule": "constant",
+    "assistant_confidence_threshold": 0.4,
+    "prompt_lookup_num_tokens": null,
+    "max_matching_ngram_size": null,
+    "assistant_early_exit": null,
+    "assistant_lookbehind": 10,
+    "target_lookbehind": 10,
+    "disable_compile": false,
+    "generation_kwargs": {},
+    "_from_model_config": false,
+    "transformers_version": "4.51.3"
+  },
+  "decode_functions": [
+    "decode_tokens",
+    "decode_parameters",
+    "decode_responses",
+    "decode_layers",
+    "decode_neurons",
+    "decode_tensors",
+    "decode_architecture",
+    "decode_fused_tensor_func"
+  ],
+  "chat_template": "{{- bos_token }}\n{%- if custom_tools is defined %}\n    {%- set tools = custom_tools %}\n{%- endif %}\n{%- if not tools_in_user_message is defined %}\n    {%- set tools_in_user_message = true %}\n{%- endif %}\n{%- if not date_string is defined %}\n    {%- if strftime_now is defined %}\n        {%- set date_string = strftime_now(\"%d %b %Y\") %}\n    {%- else %}\n        {%- set date_string = \"26 Jul 2025\" %}\n    {%- endif %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n\n{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content']|trim %}\n    {%- set messages = messages[1:] %}\n{%- else %}\n    {%- set system_message = \"\" %}\n{%- endif %}\n\n{{- \"<|start_header_id|>system<|end_header_id|>\n\n\" }}\n{%- if tools is not none %}\n    {{- \"Environment: ipython\n\" }}\n{%- endif %}\n{{- \"Cutting Knowledge Date: December 2025\n\" }}\n{{- \"Today Date: \" + date_string + \"\n\n\" }}\n{%- if tools is not none and not tools_in_user_message %}\n    {{- \"You have access to the following functions. To call a function, please respond with JSON for a function call.\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value).\"' }}\n    {{- \"Do not use variables.\n\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\n\n\" }}\n    {%- endfor %}\n{%- endif %}\n{{- system_message }}\n{{- \"<|eot_id|>\" }}\n\n{%- if tools_in_user_message and not tools is none %}\n    {%- if messages | length != 0 %}\n        {%- set first_user_message = messages[0]['content']|trim %}\n        {%- set messages = messages[1:] %}\n    {%- else %}\n        {{- raise_exception(\"Cannot put tools in the first user message when there's no first user message!\") }}\n{%- endif %}\n    {{- '<|start_header_id|>user<|end_header_id|>\n\n' -}}\n    {{- \"Given the following functions, please respond with a JSON for a function call \" }}\n    {{- \"with its proper arguments that best answers the given prompt.\n\n\" }}\n    {{- 'Respond in the format {\"name\": function name, \"parameters\": dictionary of argument name and its value).\"' }}\n    {{- \"Do not use variables.\n\n\" }}\n    {%- for t in tools %}\n        {{- t | tojson(indent=4) }}\n        {{- \"\n\n\" }}\n    {%- endfor %}\n    {{- first_user_message + \"<|eot_id|>\"}}\n{%- endif %}\n\n{%- for message in messages %}\n    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}\n        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' }}\n    {%- elif 'tool_calls' in message %}\n        {%- if not message.tool_calls|length == 1 %}\n            {{- raise_exception(\"This model only supports single tool-calls at once!\") }}\n        {%- endif %}\n        {%- set tool_call = message.tool_calls[0].function %}\n        {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' -}}\n        {{- '{\"name\": \"' + tool_call.name + '\", ' }}\n        {{- '\"parameters\": ' }}\n        {{- tool_call.arguments | tojson }}\n        {{- \"}\" }}\n        {{- \"<|eot_id|>\" }}\n    {%- elif message.role == \"tool\" or message.role == \"ipython\" %}\n        {{- \"<|start_header_id|>ipython<|end_header_id|>\n\n\" }}\n        {%- if message.content is mapping or message.content is iterable %}\n            {{- message.content | tojson }}\n        {%- else %}\n            {{- message.content }}\n        {%- endif %}\n        {{- \"<|eot_id|>\" }}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' }}\n{%- endif %}\n",
+  "_commit_hash": "56e0e89a363e1508756f8784becf436653b4f9ad",
+  "auto_map": {
+    "AutoModelForCausalLM": "modeling_custom.CustomLlamaForCausalLM"
+  },
+  "conceptual_features": {
+    "grouping_logic": true,
+    "reward_alignment": true,
+    "reasoning_tuned": true,
+    "group_size": 128
+  },
+  "safetensors": true
+}

generation_config.json

@@ -0,0 +1,9 @@
+{
+  "do_sample": true,
+  "max_new_tokens": 100,
+  "no_repeat_ngram_size": 3,
+  "repetition_penalty": 1.2,
+  "temperature": 0.7,
+  "top_p": 0.9,
+  "transformers_version": "4.51.3"
+}

modeling_custom.py

@@ -0,0 +1,12 @@
+
+from transformers.models.llama.modeling_llama import LlamaForCausalLM, LlamaConfig
+import torch
+from transformers.utils import logging
+
+logger = logging.get_logger(__name__)
+
+class CustomLlamaForCausalLM(LlamaForCausalLM):
+    def __init__(self, config: LlamaConfig):
+        super().__init__(config)
+        logger.info("CustomLlamaForCausalLM initialized.")
+