Update README.md

README.md

@@ -2,73 +2,295 @@
 license: apache-2.0
 quantized_by: bartowski
 pipeline_tag: text-generation
+lm_studio:
+  param_count: 7b
+  use_case: general
+  release_date: 15-04-2024
+  model_creator: microsoft
+  prompt_template: vicuna
+  system_prompt: A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.
+  base_model: mistral
+  original_repo: microsoft/WizardLM-2-7B
 ---
 
+## 💫 Community Model> WizardLM-2-7B by Microsoft
+
 ## Llamacpp Quantizations of WizardLM-2-7B
 
-Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b2675">b2675</a> for quantization.
+*👾 [LM Studio](https://lmstudio.ai) Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on [Discord](https://discord.gg/aPQfnNkxGC)*.
+
+**Model creator:** [Microsoft](https://huggingface.co/microsoft)<br>
+**Original model**: [WizardLM-2-7B](https://huggingface.co/microsoft/WizardLM-2-7B)<br>
+**GGUF quantization:** provided by [bartowski](https://huggingface.co/bartowski) based on `llama.cpp` release [b2675](https://github.com/ggerganov/llama.cpp/releases/tag/b2675)<br>
+
+## Model Summary:
+
+WizardLM 2 7B is a followup model to the original and highly successful WizardLM line of models. This model is trained to excel at multi-turn conversations, and does so very successfully, outclassing models more than twice its size.<br>
+This model should be used for general conversation and world knowledge, but as with most models these days will be relatively competent at coding and reasoning as well.
+
+## Prompt Template:
 
-Original model: https://huggingface.co/microsoft/WizardLM-2-7B
+For now, you'll need to make your own template. Choose the `LM Studio Blank Preset` in your LM Studio.
 
-All quants made using imatrix option with dataset provided by Kalomaze [here](https://github.com/ggerganov/llama.cpp/discussions/5263#discussioncomment-8395384)
+Then, set the system prompt to whatever you'd like (check the recommended one below), and set the following values:<br>
+`System Message Suffix`: '\n\n'<br>
+`User Message Prefix`: 'USER:'<br>
+`User Message Suffix`: '\nASSISTANT:'
 
-## Prompt format
+Under the hood, the model will see a prompt that's formatted like so:
 
 ```
-{system_prompt}
-USER: {prompt}
+A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.
+USER:{prompt}
 ASSISTANT: </s>
 ```
 
-## Download a file (not the whole branch) from below:
+## Use case and examples
+
+WizardLM 2 was tuned for improved performance on complex chat, multilingual, reasoning and agent tasks. This makes it a great model to use when wanting to chat back and forth and have reasoning based discussions.
+
+### World knowledge:
+
+```
+A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.
+USER:How far away should walnut trees be planted from one another?
+ASSISTANT:The spacing of walnut trees (Juglans spp.) depends on several factors, including the specific variety, intended use (commercial orchard, nut production, shade tree, etc.), and available space. Additionally, walnuts are allelopathic, meaning they produce compounds like juglone that can inhibit the growth of nearby plants. Here are some general guidelines:
+
+1. **Commercial Orchards**: For commercial nut production, walnut trees are typically planted 20 to 30 feet apart within the rows and about 40 to 60 feet between the centers of adjacent rows, depending on the variety and local climate conditions. This spacing allows for adequate air circulation, light penetration, and space for machinery access.
+
+2. **Nut and Fruit Production**: If you're planting walnut trees primarily for nut or fruit production (even if it's a small-scale operation), you'll want to ensure they have enough room to develop fully without competition from neighboring trees. A spacing of 20 to 30 feet within the rows and 50 to 60 feet between rows is common.
+
+3. **Shade Trees**: If you're planting walnut trees as shade trees, you can space them further apart—perhaps 40 to 60 feet center to center. This will allow for sufficient growth without overcrowding and will provide ample space for the trees to mature with plenty of sunlight reaching the ground.
+
+4. **Hedgerows or Windbreaks**: For creating a living hedgerow or windbreak, walnuts can be planted more closely, around 10 to 20 feet apart, depending on how dense you want the barrier to be.
+
+5. **Considerations for Allelopathy**: To minimize the impact of juglone, it's best to avoid planting sensitive species near walnut trees. If you're planting a diverse mixed-nut or fruit orchard, consider including species that are not affected by juglone or plant them well outside the root zone of the walnuts.
+
+6. **Soil Health**: Proper soil management and occasional soil amendments can help mitigate allelopathic effects and improve plant growth around walnut trees.
+
+7. **Local Climate and Soil**: Always take into account local climate conditions, soil type, and drainage when determining spacing. Trees planted too close together in less-than-ideal conditions can suffer from overcrowding, disease, and pest issues.
+
+8. **Future Growth**: Remember that walnut trees can grow quite large, with some species reaching heights of 70 feet or more at maturity. Ensure that the spacing allows for adequate clearance for growth without interfering with buildings, power lines, or other obstacles.
+
+Ultimately, the best spacing for your walnut trees will depend on your specific goals and conditions. Consult with local extension services, arborists, or nurseries for advice tailored to your situation.
+USER:Can you elaborate on what allelopathy and juglone are? What's the risks to sensitive plants?
+ASSISTANT:Allelopathy is a biological interaction where one organism releases chemicals into its environment that affect the growth, survival, or reproduction of other organisms—often other plants. These chemical compounds can be beneficial (as a defense mechanism against herbivores or competitors), neutral, or detrimental to surrounding flora and fauna.
+
+**Juglone (5-hydroxy-1,4-naphthoquinone)**, specifically, is one such allelochemical produced by walnut trees (and other Juglans species). It's the primary active constituent of juglone and is responsible for its strong antimicrobial, antifungal, and insecticidal properties. However, it's most well-known for its phytotoxic effects on certain plants.
+
+**Risks to Sensitive Plants:**
+
+1. **Root Death**: Juglone can cause root injury or death in many plants, which may not be immediately apparent since the toxic effect is systemic and can occur without visible signs on above-ground parts.
+
+2. **Germination Inhibition**: Seeds of sensitive species planted near walnut trees may fail to germinate due to juglone's presence in the soil.
+
+3. **Stunted Growth**: Plants that do manage to germinate and grow near walnut trees may exhibit stunted growth or yellowing leaves, indicating stress from juglone exposure.
+
+4. **Decline and Death**: Long-term exposure can lead to the decline of sensitive plants, eventually resulting in their death.
+
+5. **Soil Concentration**: Juglone concentrates in the soil over time, especially in organic matter like compost or mulch. This means that the risk of allelopathic effects can increase with the decomposition of walnut leaves, twigs, and other parts.
+
+6. **Variety of Affected Plants**: The sensitivity to juglone varies among plant species. Some plants are highly sensitive, while others show little to no effect. Commonly affected plants include tomatoes, potatoes, peas, and some ornamental species like hostas and daylilies. However, the degree of sensitivity can vary even within a plant species.
+
+**Mitigation Strategies:**
+
+- **Spatial Separation**: Plant sensitive species well away from walnut trees to minimize the risk of juglone exposure. A minimum distance of 100 feet is often recommended, but this can vary depending on factors like soil type and local conditions.
+
+- **Soil Management**: Avoid using walnut leaf mulch or compost in beds with sensitive plants. If you're unsure about the contents of compost or soil amendments, it's best to avoid them near walnut trees.
+
+- **Barrier Planting**: Use a buffer zone with non-sensitive plant species to act as a barrier between the walnut trees and sensitive plants.
+
+- **Monitoring**: Regularly monitor the health of plants in or near walnut tree canopies for signs of stress or damage.
+
+- **Professional Advice**: Consult with local extension services or arborists for specific advice tailored to your region and garden layout.
+
+It's important to note that while juglone is a well-documented allelochemical, the impact on neighboring plants can be variable, and not all plants are equally affected. Research on your particular plant species of concern can help determine the level of risk.</s>
+```
+
+## Conversational:
 
-| Filename | Quant type | File Size | Description |
-| -------- | ---------- | --------- | ----------- |
-| [WizardLM-2-7B-Q8_0.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-Q8_0.gguf) | Q8_0 | 7.69GB | Extremely high quality, generally unneeded but max available quant. |
-| [WizardLM-2-7B-Q6_K.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-Q6_K.gguf) | Q6_K | 5.94GB | Very high quality, near perfect, *recommended*. |
-| [WizardLM-2-7B-Q5_K_M.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-Q5_K_M.gguf) | Q5_K_M | 5.13GB | High quality, *recommended*. |
-| [WizardLM-2-7B-Q5_K_S.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-Q5_K_S.gguf) | Q5_K_S | 4.99GB | High quality, *recommended*. |
-| [WizardLM-2-7B-Q4_K_M.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-Q4_K_M.gguf) | Q4_K_M | 4.36GB | Good quality, uses about 4.83 bits per weight, *recommended*. |
-| [WizardLM-2-7B-Q4_K_S.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-Q4_K_S.gguf) | Q4_K_S | 4.14GB | Slightly lower quality with more space savings, *recommended*. |
-| [WizardLM-2-7B-IQ4_NL.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-IQ4_NL.gguf) | IQ4_NL | 4.12GB | Decent quality, slightly smaller than Q4_K_S with similar performance *recommended*. |
-| [WizardLM-2-7B-IQ4_XS.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-IQ4_XS.gguf) | IQ4_XS | 3.90GB | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
-| [WizardLM-2-7B-Q3_K_L.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-Q3_K_L.gguf) | Q3_K_L | 3.82GB | Lower quality but usable, good for low RAM availability. |
-| [WizardLM-2-7B-Q3_K_M.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-Q3_K_M.gguf) | Q3_K_M | 3.51GB | Even lower quality. |
-| [WizardLM-2-7B-IQ3_M.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-IQ3_M.gguf) | IQ3_M | 3.28GB | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
-| [WizardLM-2-7B-IQ3_S.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-IQ3_S.gguf) | IQ3_S | 3.18GB | Lower quality, new method with decent performance, recommended over Q3_K_S quant, same size with better performance. |
-| [WizardLM-2-7B-Q3_K_S.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-Q3_K_S.gguf) | Q3_K_S | 3.16GB | Low quality, not recommended. |
-| [WizardLM-2-7B-IQ3_XS.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-IQ3_XS.gguf) | IQ3_XS | 3.01GB | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
-| [WizardLM-2-7B-IQ3_XXS.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-IQ3_XXS.gguf) | IQ3_XXS | 2.82GB | Lower quality, new method with decent performance, comparable to Q3 quants. |
-| [WizardLM-2-7B-Q2_K.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-Q2_K.gguf) | Q2_K | 2.71GB | Very low quality but surprisingly usable. |
-| [WizardLM-2-7B-IQ2_M.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-IQ2_M.gguf) | IQ2_M | 2.50GB | Very low quality, uses SOTA techniques to also be surprisingly usable. |
-| [WizardLM-2-7B-IQ2_S.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-IQ2_S.gguf) | IQ2_S | 2.31GB | Very low quality, uses SOTA techniques to be usable. |
-| [WizardLM-2-7B-IQ2_XS.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-IQ2_XS.gguf) | IQ2_XS | 2.19GB | Very low quality, uses SOTA techniques to be usable. |
-| [WizardLM-2-7B-IQ2_XXS.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-IQ2_XXS.gguf) | IQ2_XXS | 1.99GB | Lower quality, uses SOTA techniques to be usable. |
-| [WizardLM-2-7B-IQ1_M.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-IQ1_M.gguf) | IQ1_M | 1.75GB | Extremely low quality, *not* recommended. |
-| [WizardLM-2-7B-IQ1_S.gguf](https://huggingface.co/bartowski/WizardLM-2-7B-GGUF/blob/main/WizardLM-2-7B-IQ1_S.gguf) | IQ1_S | 1.61GB | Extremely low quality, *not* recommended. |
+```
+A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.
+USER:Give me 3 fun, inexpensive summer time activities I can plan with friends
+ASSISTANT:Absolutely! Summertime is perfect for a variety of fun and budget-friendly activities with friends. Here are three ideas that are sure to create memorable experiences without breaking the bank:
+
+1. **Backyard Movie Night:**
+   - **Setup:** Transform your backyard into a cinema by hanging a white sheet on a fence or installing an outdoor projector screen. Use a projector to play your chosen movie, or even a large screen TV if you have one that can handle the outdoors.
+   - **Comfort:** Set up some blankets, cushions, and bean bags for seating. Don't forget some pillows and throws for extra comfort.
+   - **Snacks:** Create a concession stand with inexpensive movie treats like popcorn, homemade candy, and DIY ice cream sundaes.
+   - **Decor:** Use string lights or solar-powered pathway lights to set a moody ambiance. You can also hand paper lanterns for a warm glow.
+   - **Selection:** Choose a genre that everyone loves or take votes on the movies to ensure everyone's happy.
+
+2. **DIY Scavenger Hunt:**
+   - **Planning:** Design a scavenger hunt with clues leading to different locations, both indoors and outdoors. These can be riddles, trivia questions, or puzzles that lead participants from one spot to the next.
+   - **Prizes:** Keep costs low by offering small, fun prizes at the end, like homemade treats, a playlist of favorite songs, or a small trinket from a dollar store.
+   - **Teams:** Divide into teams for a more competitive edge, which can also help manage group size and maintain excitement throughout the hunt.
+   - **Variations:** You can tailor the scavenger hunt to various themes, such as historical facts about your neighborhood, nature-themed (finding specific types of plants or animals), or even a book-inspired treasure hunt if everyone shares a favorite book.
+
+3. **Outdoor Adventure Day:**
+   - **Hiking or Biking:** Find local trails for hiking or biking. Pack a backpack with water, snacks, and a first aid kit. Don't forget sunscreen and bug spray!
+   - **Picnic:** Plan a picnic at a scenic spot you'll encounter during your hike or bike ride. Prepare simple meals like sandwiches, fruit salads, and homemade cookies. Keep it potluck style if everyone wants to contribute.
+   - **Water Activities:** If there's access to water, bring along inflatable toys, kayaks, or canoes for a fun time on the water. A local lake or river can offer a refreshing break from the heat.
+   - **Exploration:** Bring along a camera or smartphone to document your adventure. Take photos of unique rocks, flowers, or wildlife you encounter.
+   - **Games:** Include lawn games like frisbee, bocce ball, or a portable cornhole set for some friendly competition.
+
+These activities encourage bonding, physical activity, and provide a great escape from the routine of daily life. They're also adaptable to different interests and can be scaled up or down depending on your group size and preferences. Enjoy your summer with friends!
+USER:I love the idea of a DIY scavenger hunt.. But my friends really like challenges, would you be able to get me started with 5 example scavenger hunt locations and clever hints for finding them?
+ASSISTANT:Absolutely! A challenge-oriented scavenger hunt can be a blast. Here are five example locations with clever hints for each. These hints are designed to be riddles or puzzles that encourage creative thinking and problem-solving. Adjust the complexity as needed based on your friends' preferences.
+
+1. **Local Park:**
+   - *Hint:* "Beneath the watchful eyes of those who converse in rhymes, rest upon a bench where poets often chimes." This clue refers to a park with a statue or sculpture depicting literary figures (like Shakespeare or Shel Silverstein), suggesting that the next clue is on one of the benches near this statue.
+
+2. **Historic Landmark:**
+   - *Hint:* "Where history's whispers dance through time, stand by the pillar where echoes chime." This clue points to a specific pillar or marker at a historic landmark, perhaps an old fort, lighthouse, or monument, where the next clue is hidden or attached.
+
+3. **Near a Mural or Public Art Installation:**
+   - *Hint:* "Amidst the colors that tell a city's tale, find your next clue where art has set its sail." This hint directs participants to a mural or large-scale public artwork where the next clue might be attached to or hidden within the artwork itself.
+
+4. **At a Library or Bookstore:**
+   - *Hint:* "In the house of stories, where knowledge reigns supreme, seek the tome that's marked with a poet's dream." This clue suggests searching for a book with a poem inside it on a shelf within the library or bookstore. The next clue could be written in the margin of the page or tucked into the book.
+
+5. **By a Natural Feature, Like a Waterfall or Unique Tree:**
+   - *Hint:* "Where water falls in a cascade of liquid silver, find your next clue by nature's giver." This hint refers to a natural waterfall or similar water feature. The next clue might be hidden at the base of the feature, on a rock above the water, or along the trail leading to it.
+
+To expand the hunt:
+- Start at a central location and assign a point value to each clue based on difficulty or distance.
+- Include a variety of clue types, such as riddles, directional clues (like "north of the fountain"), or physical challenges (like "perform a dance move" at a certain spot).
+- Have a mix of individual and group tasks to encourage teamwork.
+- Consider incorporating QR codes at each location that link to a video clue or a virtual challenge.
+
+Remember to ensure that all participants have access to the clues, either by printing them out or sharing them electronically. Also, make sure to plan the hunt so that it's safe and respectful of private property and local regulations. Have fun crafting your scavenger hunt adventure!</s>
+```
+
+## Coding:
+
+```
+A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.
+USER:Using Python, what's the most efficient way to check if a substring appears more than once in a given string?
+ASSISTANT:To efficiently check if a substring appears more than once in a given string in Python, you can use a combination of data structures and algorithms. Here are a few methods with increasing efficiency:
+
+1. **Simple Iteration (Naive Approach)**:
+
+    def has_substring_more_than_once(s, sub):
+        if sub.len() == 0:
+            return True
+        substr_count = 0
+        start = s.find(sub)
+        while start != -1:
+            substr_count += 1
+            start = s.find(sub, start + 1)
+        return substr_count > 1
 
-## Which file should I choose?
+This method is straightforward but has a time complexity of O(n^2) for checking the presence of the substring multiple times in the string.
+
+2. **Using `str.count()`**:
+
+    def has_substring_more_than_once(s, sub):
+        return s.count(sub) > 1
+
+This method is more Pythonic and uses an internal counting mechanism, but it still has a time complexity of O(n) for counting the occurrences.
+
+3. **Leveraging `str.find()` with a counter**:
+
+    def substring_occurrences(s, sub):
+        count = 0
+        pos = s.find(sub)
+        while pos != -1:
+            count += 1
+            pos = s.find(sub, pos + 1)
+        return count > 1
+
+This method improves upon the first one by using `str.find()` which is an efficient way to locate substrings. It still has a time complexity of O(n^2), but it's more practical for small to medium-sized strings.
+
+4. **Using Regular Expressions**:
+
+    import re
+    
+    def has_substring_more_than_once(s, sub):
+        return len(re.findall(re.escape(sub), s)) > 1
+
+This method uses the `re` module to search for all occurrences of the substring (escaped to handle special regex characters) and returns `True` if there are more than one match. The time complexity is generally O(n) where n is the length of the string, but it can be slower if the regex engine performs a lot of backtracking for certain patterns.
+
+5. **Efficient Scanning with KMP (Knuth-Morris-Pratt) Algorithm**:
+For an even more efficient approach, you can implement or use a library that implements the KMP algorithm, which preprocesses the substring to create a "lazy" search pattern that skips characters when a mismatch occurs. This reduces the average time complexity from O(n*m) (where n is the length of the string and m is the length of the substring) to O(n+m).
+
+Here's an example implementation of KMP:
+
+    def kmp_search(pattern, text):
+        """
+        Returns the starting positions in text where pattern occurs.
+        """
+        def compute_lps(psz):
+            length = 0
+            lps = [0] * len(psz)
+            i = 1
+            length = 0
+            while i < len(psz):
+                if psz[i] == psz[length]:
+                    length += 1
+                    lps[i] = length
+                    i += 1
+                else:
+                    if length != 0:
+                        length = lps[length - 1]
+                    else:
+                        lps[i] = 0
+                        i += 1
+            return lps
+    
+        lps = compute_lps(pattern)
+        i = j = 0
+        occurrences = []
+        while i < len(text):
+            if text[i] == pattern[j]:
+                i += 1
+                j += 1
+                if j == len(pattern):
+                    occurrences.append(i - j)
+                    j = lps[j - 1]
+            elif j > 0:
+                j = lps[j - 1]
+            else:
+                i += 1
+        return occurrences
+    
+    def has_substring_more_than_once(s, sub):
+        positions = kmp_search(sub, s)
+        return len(positions) > 1
+
+6. **Using `index` and `rindex` for Exact Positions**:
+If you only need to know if the substring appears more than once without the exact positions, you can use `index` and `rindex` which are faster than `find` because they do not search the entire string after a match is found:
+
+    def has_substring_more_than_once(s, sub):
+        first = s.index(sub)
+        last = s.rindex(sub)
+        return first != last
+
+This method assumes that the substring appears at the beginning and at the end of the string, which is a special case but can be efficient if true.
+
+For general use, the KMP algorithm or regular expressions with `re.findall()` are likely your best bet for efficiency in most cases.</s>
+```
 
-A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9)
+## Technical Details
 
-The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.
+WizardLM 2 applies several new methods for training compared to the original iteration, truly showing just how much the Open Source AI world has advanced since their intial offerings.
 
-If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.
+The first of which is Progress Learning. Rather than applying all training data at once, the team applied stage-by-stage training by partitioning the data into multiple sections and training on each one after the other.
 
-If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.
+AI Align AI (AAA) is another new process, whereby various state-of-the-art LLMs are allowed to co-teach and improve from each other, using simulated chats, quality judging, and improvement suggestions. They also participate in self-teaching in a similar manor.
 
-Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.
+The model then underwent Supervised Learning, Stage-DPO, and Evol-Instruct and Instruction&Process Supervised Reinforcement Learning (RLEIF) which uses an instruction quality reward model and a supervision reward model for more precise correctness.
 
-If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.
+The results are a model that performs exceptionally well on the automatic MT-Bench evaluation. 
 
-If you want to get more into the weeds, you can check out this extremely useful feature chart:
+For more information, check the WizardLM2 blog post [here](https://wizardlm.github.io/WizardLM2/)
 
-[llama.cpp feature matrix](https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix)
+## Special thanks
 
-But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.
+🙏 Special thanks to [Georgi Gerganov](https://github.com/ggerganov) and the whole team working on [llama.cpp](https://github.com/ggerganov/llama.cpp/) for making all of this possible.
 
-These I-quants can also be used on CPU and Apple Metal, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.
+🙏 Special thanks to [Kalomaze](https://github.com/kalomaze) for his dataset (linked [here](https://github.com/ggerganov/llama.cpp/discussions/5263)) that was used for calculating the imatrix for these quants, which improves the overall quality!
 
-The I-quants are *not* compatible with Vulcan, which is also AMD, so if you have an AMD card double check if you're using the rocBLAS build or the Vulcan build. At the time of writing this, LM Studio has a preview with ROCm support, and other inference engines have specific builds for ROCm.
+## Disclaimers
 
-Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski
+LM Studio is not the creator, originator, or owner of any Model featured in the Community Model Program. Each Community Model is created and provided by third parties. LM Studio does not endorse, support, represent or guarantee the completeness, truthfulness, accuracy, or reliability of any Community Model.  You understand that Community Models can produce content that might be offensive, harmful, inaccurate or otherwise inappropriate, or deceptive. Each Community Model is the sole responsibility of the person or entity who originated such Model. LM Studio may not monitor or control the Community Models and cannot, and does not, take responsibility for any such Model. LM Studio disclaims all warranties or guarantees about the accuracy, reliability or benefits of the Community Models.  LM Studio further disclaims any warranty that the Community Model will meet your requirements, be secure, uninterrupted or available at any time or location, or error-free, viruses-free, or that any errors will be corrected, or otherwise. You will be solely responsible for any damage resulting from your use of or access to the Community Models, your downloading of any Community Model, or use of any other Community Model provided by or through LM Studio.