BMPixel

The assistant will play the role of an advanced reasoning-based problem-solving expert. The assistant use an *contemplator* technique to make extremely detailed and advanced reasoning over problems. Follow these steps to structure your contemplating small talk. 

- **Step-by-step reasoning**: Start with the problem and **break it down**, analyzing each detail. Clearly explain your reasoning at each step. Propose hypotheses and test them. If one fails (It's very likely), adjust it and keep exploring. **Break complex steps into simpler ones** to make the process easier to manage.

- **Thought jumps**: If new ideas arise, revisit earlier steps and explain why. When trying different approaches, note why previous hypotheses failed.

- **Heavy self doubt**: Always assume previous steps containing flaws. Always try the best to spot errors in previous reasonings. NEVER BLINDLY AGREE ON PREVIOUS REASONINGS. If a potential solution is found, try your hardest to negate it.

- **Tree-like path search**: Contemplating pro

BMPixel

BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Assistant Generated//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:Asia/Hong_Kong
BEGIN:STANDARD
DTSTART:19700101T000000
TZOFFSETFROM:+0800
TZOFFSETTO:+0800

BMPixel

BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Assistant Generated//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:Asia/Hong_Kong
BEGIN:STANDARD
DTSTART:19700101T000000
TZOFFSETFROM:+0800
TZOFFSETTO:+0800

BMPixel

{
    // Remove titlebar
    "apc.header": {
        "height": 37
    },
    "apc.sidebar.titlebar": {
        "height": 37
    },
    "window.titleBarStyle": "native",
    // Transparency

BMPixel

# 主要改进
# - 换成 Full finetuning
# Model define
base_model: /cephfs/panwenbo/work/models/Faro-34B
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer

# is_qwen_derived_model: true
trust_remote_code: true
chat_template: chatml

BMPixel

# Model define
base_model: 01-ai/Yi-9B-200K
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer

trust_remote_code: true
chat_template: chatml

# Data
datasets:

BMPixel

# ============================================
# Zsh Configuration with better default
# Zsh configs that anyone should be welcomed to incorporate
# No complex prompts, just the essentials:
# - No aliases and keybindings.
# - No need for third-party plugins.
# ============================================

# ========= Completion & History Setup =========
autoload -Uz compinit

BMPixel

from collections import Counter
import pickle
from transformers import PreTrainedTokenizerFast
import json


# Load the base tokenizer from the model
base_tokenizer = PreTrainedTokenizerFast.from_pretrained("/cephfs/panwenbo/work/models/Meta-Llama-3-8B")

BMPixel

# 主要改进
# - 换成 Full finetuning
# Model define
base_model: /cephfs/panwenbo/work/models/Faro-34B
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer

# is_qwen_derived_model: true
trust_remote_code: true
chat_template: chatml

BMPixel

# 主要改进
# - 换成 Full finetuning
# Model define
base_model: /cephfs/panwenbo/work/models/Yi-9B-200K
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer

# is_qwen_derived_model: true
trust_remote_code: true
chat_template: chatml