Aether-492B-chat

chat_template.jinja

@@ -0,0 +1 @@
+{%- if tools %}{{- '<|im_start|>system\nYou are AetherMind, a helpful, honest and safe multilingual assistant.\nYou have access to the following tools (JSON schema):\n' + tools | tojson + '\nWhen you decide to call a tool, output one JSON object between <|tool_start|> and <|tool_end|>. After receiving tool results, answer the user.<|im_end|>\n' }}{%- endif %}{%- for message in messages %}{%- if message.role == 'assistant' and message.tool_calls %}{{- '<|im_start|>assistant\n' + (message.content or '') }}{%- for tc in message.tool_calls %}{{- '<|tool_start|>' + {'name': tc['function']['name'], 'arguments': tc['function']['arguments']} | tojson + '<|tool_end|>' }}{%- endfor %}{{- '<|im_end|>\n' }}{%- else %}{{- '<|im_start|>' + message['role'] + '\n' + (message['content'] or '') + '<|im_end|>\n' }}{%- endif %}{%- endfor %}{%- if add_generation_prompt %}{{- '<|im_start|>assistant\n' }}{%- endif %}

chat_template.json

@@ -0,0 +1,3 @@
+{
+  "chat_template": "{%- if tools %}{{- '<|im_start|>system\\nYou are AetherMind, a helpful, honest and safe multilingual assistant.\\nYou have access to the following tools (JSON schema):\\n' + tools | tojson + '\\nWhen you decide to call a tool, output one JSON object between <|tool_start|> and <|tool_end|>. After receiving tool results, answer the user.<|im_end|>\\n' }}{%- endif %}{%- for message in messages %}{%- if message.role == 'assistant' and message.tool_calls %}{{- '<|im_start|>assistant\\n' + (message.content or '') }}{%- for tc in message.tool_calls %}{{- '<|tool_start|>' + {'name': tc['function']['name'], 'arguments': tc['function']['arguments']} | tojson + '<|tool_end|>' }}{%- endfor %}{{- '<|im_end|>\\n' }}{%- else %}{{- '<|im_start|>' + message['role'] + '\\n' + (message['content'] or '') + '<|im_end|>\\n' }}{%- endif %}{%- endfor %}{%- if add_generation_prompt %}{{- '<|im_start|>assistant\\n' }}{%- endif %}"
+}

config.json

@@ -0,0 +1,42 @@
+{
+  "architectures": [
+    "AetherMindForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 2,
+  "dtype": "float32",
+  "eos_token_id": 3,
+  "head_dim": 32,
+  "hidden_act": "silu",
+  "hidden_size": 256,
+  "initializer_range": 0.02,
+  "intermediate_size": 768,
+  "max_position_embeddings": 1024,
+  "model_type": "aethermind",
+  "moe_aux_loss_coeff": 0.01,
+  "moe_intermediate_size": 768,
+  "moe_layers_freq": 1,
+  "num_attention_heads": 8,
+  "num_experts": 8,
+  "num_experts_per_tok": 2,
+  "num_hidden_layers": 6,
+  "num_key_value_heads": 4,
+  "pad_token_id": 0,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "rope_theta": 10000.0,
+    "rope_type": "default"
+  },
+  "rope_theta": 10000.0,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.17.0",
+  "use_cache": true,
+  "use_moe": false,
+  "vocab_size": 4326,
+  "auto_map": {
+    "AutoConfig": "configuration_aethermind.AetherMindConfig",
+    "AutoModel": "modeling_aethermind.AetherMindModel",
+    "AutoModelForCausalLM": "modeling_aethermind.AetherMindForCausalLM"
+  }
+}

generation_config.json

@@ -0,0 +1,8 @@
+{
+  "bos_token_id": 2,
+  "do_sample": false,
+  "eos_token_id": 3,
+  "max_new_tokens": 64,
+  "pad_token_id": 0,
+  "transformers_version": "5.17.0"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c08e260e255960a1a1ea2ee8b37c46c106168b73a600f2ba73b1b0038ed8a03b
+size 23323560

special_tokens_map.json

@@ -0,0 +1,12 @@
+{
+  "bos_token": "<|im_start|>",
+  "eos_token": "<|im_end|>",
+  "pad_token": "<|pad|>",
+  "unk_token": "<|unk|>",
+  "additional_special_tokens": [
+    "<|tool_start|>",
+    "<|tool_end|>",
+    "<|reasoning_start|>",
+    "<|reasoning_end|>"
+  ]
+}

tokenizer_config.json

@@ -0,0 +1,11 @@
+{
+  "backend": "tokenizers",
+  "bos_token": "<|im_start|>",
+  "eos_token": "<|im_end|>",
+  "is_local": true,
+  "local_files_only": false,
+  "model_max_length": 1024,
+  "pad_token": "<|pad|>",
+  "tokenizer_class": "TokenizersBackend",
+  "unk_token": "<|unk|>"
+}