Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+LFM2.5-1.2B-Thinking-F16.gguf filter=lfs diff=lfs merge=lfs -text

LFM2.5-1.2B-Thinking-F16.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77e8ae7d12ff42393ac81150216a24ed9dfe03616d1eb10b1edda6b886d5f191
+size 2343326400

README.md

@@ -0,0 +1,48 @@
+# LFM2.5-1.2B-Thinking-Financial-Analyst (LFM2.5 1.2B 金融分析专家版)
+
+## Overview | 概述
+This model is a specialized version of the **Liquid LFM2.5-1.2B-Thinking** model, fine-tuned to act as a professional **Financial Analyst**. It is specifically optimized for analyzing **Chinese A-share individual stocks**, interpreting **CFA-level financial principles**, and generating structured investment logic.
+
+本模型是基于 **Liquid LFM2.5-1.2B-Thinking** 的深度微调版本，旨在打造专业的**金融分析助手**。模型针对**中国A股个股咨询**、**CFA专业财务知识**以及**结构化投资逻辑**进行了深度优化。
+
+---
+
+## What's New | 模型特性
+- **Enhanced A-Share Analysis (A股深度分析)**: Learned the specific narrative style and logic of Chinese equity research reports. 更擅长以中国证券行研报告的风格和逻辑进行个股分析。
+- **CFA Professional Knowledge (CFA专业知识支撑)**: Integrated high-quality data covering accounting standards, valuation models, and ethical frameworks from the CFA curriculum. 整合了涵盖会计准则、估值模型和CFA体系下的专业财务知识。
+- **Thinking Process (逻辑推理过程)**: Retains and refines the "Thinking" capability of the base model, providing a step-by-step logical deduction before outputting the final financial conclusion. 继承并优化了原模型的“思考”能力，在给出金融结论前进行严密的逻辑推导。
+
+## Data & Direction | 微调资料与方向
+The fine-tuning involved a vast amount of specialized financial data, moving away from general conversational AI toward a domain-specific expert:
+1.  **Chinese Equity Research (中国行研数据)**: Massive collection of A-share individual stock analyses and market commentary. 累计了大量A股个股研报及市场评论。
+2.  **CFA Knowledge Base (CFA财务知识库)**: Structured data on financial statement analysis, corporate finance, and accounting logic. 系统化的财务报表分析、公司理财及会计逻辑数据。
+3.  **Specialized Financial Topics (金融专项课题)**: Deep dives into niches like **Green Bonds** (based on 2021 data) and the impact of cross-border capital flows. 涵盖绿色债券（基于2021年数据）及跨境资金流动影响等专项课题。
+
+## Origin | 模型渊源
+- **Base Model (原模型)**: `liquidai/lfm-2.5-1.2b-thinking`.
+- **Transformation (演变)**: Transformed from a general-purpose reasoning model into a structured, data-driven financial analyst. 从通用型逻辑模型演变为结构化、数据驱动的金融领域专家。
+
+## Usage | 使用方法
+
+### Option 1: LM Studio (Recommended)
+1.  Download the `.gguf` file.
+2.  Import via `lms import` or Drag & Drop.
+3.  The model is optimized for structured financial queries.
+
+### Option 2: Transformers (Python)
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+model_id = "EricLu/LFM2.5-1.2B-Financial-Analyst-Thinking" 
+model = AutoModelForCausalLM.from_pretrained(model_id, trust_remote_code=True, device_map="cuda")
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+
+prompt = "User: 请从CFA财务分析角度，评价某A股公司的现金流质量。\n\nAssistant:"
+inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
+output = model.generate(**inputs, max_new_tokens=1024)
+print(tokenizer.decode(output[0]))
+```
+
+## Disclaimer | 免责声明
+*This model is for informational purposes only and does not constitute financial advice. Small models (1.2B) may produce hallucinations; always verify critical data.*
+*本模型仅供参考，不构成任何投资建议。1.2B量级模型可能产生幻觉，请务必核实关键数据。*

chat_template.jinja

@@ -0,0 +1,45 @@
+{{- bos_token -}}
+{%- set keep_past_thinking = keep_past_thinking | default(false) -%}
+{%- set ns = namespace(system_prompt="") -%}
+{%- if messages[0]["role"] == "system" -%}
+    {%- set ns.system_prompt = messages[0]["content"] -%}
+    {%- set messages = messages[1:] -%}
+{%- endif -%}
+{%- if tools -%}
+    {%- set ns.system_prompt = ns.system_prompt + ("\n" if ns.system_prompt else "") + "List of tools: [" -%}
+    {%- for tool in tools -%}
+        {%- if tool is not string -%}
+            {%- set tool = tool | tojson -%}
+        {%- endif -%}
+        {%- set ns.system_prompt = ns.system_prompt + tool -%}
+        {%- if not loop.last -%}
+            {%- set ns.system_prompt = ns.system_prompt + ", " -%}
+        {%- endif -%}
+    {%- endfor -%}
+    {%- set ns.system_prompt = ns.system_prompt + "]" -%}
+{%- endif -%}
+{%- if ns.system_prompt -%}
+    {{- "<|im_start|>system\n" + ns.system_prompt + "<|im_end|>\n" -}}
+{%- endif -%}
+{%- set ns.last_assistant_index = -1 -%}
+{%- for message in messages -%}
+    {%- if message["role"] == "assistant" -%}
+        {%- set ns.last_assistant_index = loop.index0 -%}
+    {%- endif -%}
+{%- endfor -%}
+{%- for message in messages -%}
+    {{- "<|im_start|>" + message["role"] + "\n" -}}
+    {%- set content = message["content"] -%}
+    {%- if content is not string -%}
+        {%- set content = content | tojson -%}
+    {%- endif -%}
+    {%- if message["role"] == "assistant" and not keep_past_thinking and loop.index0 != ns.last_assistant_index -%}
+        {%- if "</think>" in content -%}
+            {%- set content = content.split("</think>")[-1] | trim -%}
+        {%- endif -%}
+    {%- endif -%}
+    {{- content + "<|im_end|>\n" -}}
+{%- endfor -%}
+{%- if add_generation_prompt -%}
+    {{- "<|im_start|>assistant\n" -}}
+{%- endif -%}

config.json

@@ -0,0 +1,57 @@
+{
+  "architectures": [
+    "Lfm2ForCausalLM"
+  ],
+  "block_auto_adjust_ff_dim": true,
+  "block_dim": 2048,
+  "block_ff_dim": 12288,
+  "block_ffn_dim_multiplier": 1.0,
+  "block_mlp_init_scale": 1.0,
+  "block_multiple_of": 256,
+  "block_norm_eps": 1e-05,
+  "block_out_init_scale": 1.0,
+  "block_use_swiglu": true,
+  "block_use_xavier_init": true,
+  "bos_token_id": 1,
+  "conv_L_cache": 3,
+  "conv_bias": false,
+  "conv_dim": 2048,
+  "conv_use_xavier_init": true,
+  "dtype": "bfloat16",
+  "eos_token_id": 7,
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 12288,
+  "layer_types": [
+    "conv",
+    "conv",
+    "full_attention",
+    "conv",
+    "conv",
+    "full_attention",
+    "conv",
+    "conv",
+    "full_attention",
+    "conv",
+    "full_attention",
+    "conv",
+    "full_attention",
+    "conv",
+    "full_attention",
+    "conv"
+  ],
+  "max_position_embeddings": 128000,
+  "model_type": "lfm2",
+  "norm_eps": 1e-05,
+  "num_attention_heads": 32,
+  "num_heads": 32,
+  "num_hidden_layers": 16,
+  "num_key_value_heads": 8,
+  "pad_token_id": 0,
+  "rope_theta": 1000000.0,
+  "tie_embedding": true,
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_pos_enc": true,
+  "vocab_size": 65536
+}

generation_config.json

@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": [
+    7
+  ],
+  "pad_token_id": 0,
+  "transformers_version": "4.57.6"
+}

special_tokens_map.json

@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|startoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|pad|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}